(Routledge Revivals) András Mócsy - Pannonia and Upper Moesia - A History of The Middle Danube Provinces of The Roman Empire-Routledge (2014)

Routledge Routledge Revivals
Taylor & Francis Group
Pannonia and Upper Moesia
In Pannonia and Upper Moesia, first published 1974, András Mócsy surveys the Middle
Danube Provinces from the latest pre-Roman Iron Age up to the beginning of the
Great Migrations. His primary concern is to develop a general synthesis of the archae-
ological and historical researches in the Danube Basin, which lead to a more detailed
knowledge of the Roman culture of the area.
The economic and social development, town and country life, culture and religion in
the Provinces are all investigated, and the local background of the so-called Illyrian
Predominance during the third century crisis of the Roman Empire is explained, as is
the eventual breakdown of Danubian Romanisation.
This volume will appeal to students and teachers of archaeology alike, as well as to
those interested in the Roman Empire – not only the history of Rome itself, but also of
the far-flung areas which together comprised the Empire’s frontier for centuries.
This page intentionally left blank
Pannonia and Upper Moesia
A History of the Middle Danube Provinces of the Roman
Empire
András Mócsy
First published in 1974
by Routledge & Kegan Paul Ltd
This edition first published in 2014 by Routledge
2 Park Square, Milton Park, Abingdon, Oxon, OX14 4RN
and by Routledge
711 Third Avenue, New York, NY 10017
Routledge is an imprint of the Taylor & Francis Group, an informa business
© 1974 András Mócsy
The right of András Mócsy to be identified as author of this work has been asserted by him in
accordance with sections 77 and 78 of the Copyright, Designs and Patents Act 1988.
All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by
any electronic, mechanical, or other means, now known or hereafter invented, including photocopying
and recording, or in any information storage or retrieval system, without permission in writing from
the publishers.
Publisher’s Note
The publisher has gone to great lengths to ensure the quality of this reprint but points out that some
imperfections in the original copies may be apparent.
Disclaimer
The publisher has made every effort to trace copyright holders and welcomes correspondence from
those they have been unable to contact.
A Library of Congress record exists under LC control number: 75306910
ISBN 13: 978-0-415-74582-6 (hbk)

ISBN 13: 978-1-315-79767-0 (ebk)
FRONTISPIECE Late Romanofficer's gilt parade-helmet(no. I) from Berkasovo
(p. 337)
Pannonia
and Upper Moesia
A History of the Middle Danube
Provincesof the Roman Empire
Andras M6csy
Professorof Archaeology
University of Budapest
Translation edited by
SheppardFrere
LONDON AND BOSTON

ROUTLEDGE & KEGAN PAUL
First publishedin I974
by Routledge& Kegan Paul Ltd
BroadwqyHouse, 68-74 Carter Lane
London EC4VfEL and
9 Park Street
Boston,Mass. 02I08, USA
Printed in Great Britain by
RichardClay (The ChaucerPress) Ltd
Bungay,Suffolk
© Andras Moc.J)' I974
No part of this book may be reproducedin
atry form withoutpermissionfrom the
publisher, exceptfor the quotation of brief
passagesin criticism
ISBN 0 7IOO 77I4 9
Contents
Preface page xix

Chapter 1 Thracians,Illyrians and Celts I
Chapter 2 Conquestof the Danuberegion and Augustanfrontier

policy 31
Chapter J Native populationand settlement 53
Chapter 4 The Danubefrontier from Vespasianto Marcus Aurelius 80
Chapter J The first age of prosperity I 12
Chapter 6 Criseson the Danube:the rise of the Illyrican soldiery 183

Chapter 7 The secondage of prosperity: rise and collapse 2I 3
Chapter S The Danubefrontier in the late Romanperiod 266
Chapter 9 The final period of prosperity 297
Chapter10 The beginningof the Dark Age 339
Abbreviations 359
Notes 363
SelectBibliography 407
Indexes 429
Plates
Vll
Plates
FRONTISPIECE Late Roman officer's gilt parade-helmet(no. I) from Ber-

kasovo. Photo. VojvodanskiMuseum,Novi Sad
PLATE IA Greek bronze vessel from the Celtic cemetery at Szob.
HungarianNational Museum
PLATE IB Late Celtic vesselwith Latin inscription DA BIBERE from the
Romano-Celticcemeteryat Cserszegtomaj.BalatonMuseum,
Keszthely
PLATE 2 Some types of barbarian coins (obverse) (1/1). Hungarian
National Museum
P L ATE 3 Reversesof coins on Plate 2
PLATE 4A Norican-Pannonianfibulae. HungarianNational Museum
PLATE 4B The impression of an intaglio showing Victoria on a late
Celtic red-painted vessel from Aquincum (4/1). Photo.
BudapestiTiirtineti Museum
PLATE 5A Tombstone of Scerviaedus,son of Sita, slain by robbers.
Photo. Kdlmdn K6nya, Budapest
PLATE 5B Tombstone of Atta, son of Bataio, a trader from Savaria.
Photo. Savaria Museum
PLA TE 5C Tombstoneof Tiberius Satto of Cambodunum(Kempten),
veteranof legio X Geminafrom Aquincum. Photo. Budapesti
TiJrtineti Museum
PLATE 6A Tombstone of Verondacus,son of Veruicus from Torok-
balint. HungarianNational Museum. Photo. J. Kardth
IX
Plates
PLATE 6B Tombstoneof Cornelius Zosimusfrom Viminacium.
Photo. Kdlmdn Kotrya, Budapest
PLATE 6c Tombstone of Cornelius Rufus from Viminacium.
Photo. Kdlmdn Konya, Budapest
PLATE 7 The Morava valley. Photo. Kdlmdn Kotrya, Budapest
PLATE 8A The late Roman fort at Boljetin in the Djerdap.
Photo. Arheololki Institut, Belgrade
PLATE 8B View of the Djerdap. Photo. Kdlmdn Kotrya, Budapest
PLATE 9A The rock-cut Roman road in the Djerdap. Photo. Kdlmdn
Kotrya, Budapest
PLATE 9B Trajan's road-inscription(the Tabula Traiani) in the Djerdap.
PLATE lOA Tombstoneof L. Licinius Lepidus from Dozmat.
Photo. Savaria Museum
PLATE lOB Tombstone of Reginus son of Troucetissaof Trier, from
Aquincum. Photo. BudapestiTijrteneti Museum
PLATEIIA Tombstoneof C. Sextilius Senecio,decurion of Scarbantia.
PLATEIIB Tombstoneof SemproniusMarcellinus and his family from
Savaria. Photo. Savaria Museum
PLATE IIC Tombstoneof Aurelius Rufinus from Intercisa. Hungarian
National Museum
PLATE 12A Tombstoneof Aelius Munatius from Intercisa. Hungarian
National Museum
PLATE 12B Tombstone of Aelius Septimus from Brigetio. Hungarian
National Museum
PLATE 12C Tombstone of Bozi, daughter of Vellasa from Ercsi.
PLATE 13A Stone slab from an aedicula at Bolcske. Hungarian National
Museum
PLATE 13B Coins of Trajanreferringto the mines. (i) left, RIC 706, metalli
Ulpiani Pan (2/1). (ii) right, RIC 704, Dardanici (2/1). Photo.
R. L. Wilkins, by courtesyof the AshmoleanMuseum,Oxford
PLATE I4A The mountainsof the metalla Ulpiana nearUlpiana(Gracanica).
Photo. Kdlmdn Kotrya, Budapest
PLATE I4B The Danubeat Visegrad. Photo. Kdlmdn Konya, Budapest
PLATE 15A, B Finds from cart-graves.HungarianNational Museum
PLATE 16A Aquincum municipium: the mace/lum. Photo. Kdlmdn Kotrya,
Budapest
x
Plates
PLATE 16B Aquincum mUnlClplum: the 'big house'. Photo. Kdlmdn
Konya, Budapest
PLATE 17A, B Aquincum municipium: the so-called 'large baths'.
PhotosKdlmdn Konya, Budapest
PLATE 18A Carnuntummunicipium: House VI. Photo. Carnuntum
Museum
PLATE 18B Carnuntummunicipium: House IV. Photo. Carnuntum
Museum
PLATE 19A Carnuntum:the civil amphitheatre. Photo. H. Kral
PLATE 19B Carnuntum:the four-way arch (Heidentor). Photo. H. Kral
PLATE 20A Excavations at the municipium Dardanorum: the horreum.
Photo. by courtesyof the late E. Cerskov
PLATE 20B The Roman city wall of Scarbantiaas reconstructedin the
Middle Ages. Photo. Kdlmdn Konya, Budapest
PLATE 21 The Aszar hoard. Hungarian National Museum.
Photos]. Kardth
(a) Jewelry
(b) BronzeVessel I
(c) BronzeVessel 3
PLATE 22 The military diploma from the Aszar hoard. Hungarian
National Museum
PLATE 23A The Nymphaeum of the Sanctuary at Gorsium.
PLATE 23B Sarcophagusof cline type from Viminacium. Photo. Kdlmdn
Konya, Budapest
PLATE 24A Some types of Pannonian stamped pottery. Hungarian
National Museum
PLATE 24B Samianware from the 'Siscia' pottery, found at Viminacium.
P LA T E 25 A A mould of the potter Pacatus from Aquincum.
Photo. BudapestiTiJrtineti Museum
PLATE 25B The EmperorMarcus Aurelius: stamp for a potter's mould
from Aquincum. Photo. BudapestiTiirtineti Museum
PLATE 26A Tombstoneof Trebia Lucia from Timacum minus.
Photo. Kdlman Konya, Budapest
PLATE 26B Tombstone of Aelia Clementilla from Ulpiana.
PLATE 27A Scarbantia:the CapitolineTriad, Juppiter,JunoandMinerva.
xi
Plates
PLATE 27B Savaria: relief from the Iseum. Photo. Kdlmdn Kdnya,
Budapest
PLATE 28A, B Mosaicpavementsfrom Aquincum. Photo. BudapestiTiJrtCneti
Museum
PLATE 29A The sarcophagusof Pia Celerina from Aquincum.
Photo. Kdlmdn Kdnya, Budapest
PLATE 29B The sarcophagusof the scholasticusL. SeptimiusFuscusfrom
Aquincum. Photo. BudapestiTortCneti Museum
PLATE 29C The sarcophagusof the Interpres Dacorum from Brigetio.
PLATE 30A Statuette of Negro boy from Aquincum. Photo. Kdlmdn
Kdnya, Budapest
PLATE 30B Burial-vault from Brestovik, east of Singidunum.
Photo. ArheolofkiInstitut, Belgrade
PLATE3IA Stone slab with relief of the god Silvanus from Aquincum.
Photo. BudapestiTiirtCneti Museum
PLATE 3IB Relief of Diana from Balatonvilagos. Photo. Kdlmdn Kdnya,
Budapest
PLATE 32A Stone slab with mythological scene: Bellerophon and the
Chimaera.From Intercisa.HungarianNational Museum
PLATE 32B Stone slab with mythological scene: Aeneas escapingfrom
Troy. From Intercisa.HungarianNational Museum
PLATE BA Stone slab with mythological scene: Priam and Achilles.
From Aquincum. Photo. BudapestiTiirtCneti Museum
PLATE33B Stone slab with mythological scene: Achilles and Hector.
PLATE 34A Lead tablet showing the Danubian Rider-gods. Hungarian
National Museum. Photo. J. Kardth
PLATE 34B Altar erectedat Carnuntumto Mithras by the Tetrarchsin 308.
CarnuntumMuseum. Photo. H. Kral
PLATE35A Glass vas diatretum from Szekszard. Hungarian National
Museum. Photo. J. Kardth
PLATE35B Silver vessel of Licinius from Esztergom. Hungarian
National Museum
PLATE 36A Late Romanfort at Tokod. Photo. Kdlmdn K6nya, Budapest
PLATE 36B Valentinianic burgus at Visegrad. Photo. Kdlmdn Kdnya,
Budapest
PLATE 37A, B Gamzigrad: late Roman imperial palace. Photos Kdlmdn
Kdnya, Budapest
xii
Plates
PLATE 38A The early Christian basilica at Fenekpuszta. Photo. by
courtesyoj Dr K. Sdgi
PLATE 38B The early Christian basilica at Ulpianum. Photo. Kdlmdn
KOlrya, Budapest
PLATE 39A Sirmium: the late Roman baths. Photo. Kdlmdn KOlrya,
Budapest
PLATE 39B Sirmium: part of the late Roman palace. Photo. Kdlmdn
Konya, Budapest
PLATE 40A Sopianae(Pees):paintedtomb NO.2. Photo. Kdlmdn KOlrya,
Budapest
PLATE 40B Sopianae(Pees):tomb with painting of SS. Peterand Paul.
PLATE 4IA Sirmium: fragments of a cancellum (screen) from a Christian
basilica, built into a modern wall. Photo. Kdlmdn KOlrya,
Budapest
PLATE 4IB Paintedtomb in the Christian cemeteryat Jagodin Mahala.
Photo. Narodni Museum,Nil
PLATE 42A, B Details of paintedfigures in the Christiancemeteryof Jagodin
Mahala. PhotosNarodni Museum,Nil
PLATE 43A Painted figure from a tomb at Beska. Photo. Vr:jvotijanski
Museum,Novi Sad
PLATE 43B Mosaic from the Constantinian villa of Mediana.
Photo. Kdlmdn KOlrya, Budapest
PLATE 44A The augur'sstaff from Brigetio. HungarianNationalMuseum
PLATE 44B Bronze plate from late Roman box: Kisarpas. Hungarian
National Museum
PLATE 44C Bronze plates from late Roman boxes: Sagvar. Hungarian
National Museum
PLATE 45A Silver tripod from Polgardi. Hungarian National Museum.
Photo. J. Kardth
PLATE 45B Late Roman inscribed gold pin from Fenekpuszta.
xiii
Figures
Figure I The areaof the middle Danubein the fifth centuryB.C. page 3
Figure 2 The areaof the middle Danubein the fourth centuryB.C. 6
Figure j The areaof the middle Danubein the third centuryB.C. 8
Figure 4 The areaof the middle Danubein the secondcenturyB.C. II
Figure J The area of the middle Danube in the first half of the
first century B.C. 16
Figure 6 The areaof the middle Danubein the secondhalf of
the first century B.C. 20
Figure 7 The distribution of Greek,Romanrepublicanand 29
native coins
Figure 8 The tow-pathin the Djerdap (reconstructionby 46
E. Swoboda)
Figure 9 The native peoplesof Pannonia 54
Figure IO Tombstoneswith astral symbols 62
Figure I I Languagesand tribal nameswith Latin suffixes 64
Figure I2 The native peoplesof Upper Moesia 67
Figure Ij Town plans I: Emonaand Savaria 75
Figure I4 The environsof Aquincum 87
Figure I J Romanbuildings in the Barbaricum 9°
Figure I6 The division of Pannoniaunder Trajan 93
Figure I7 Inscription of Mercatorfrom Szeged 101
Figure I8 Danubeforts in Pannonia 108
Figure I9 The distribution of someItalian families in Pannonia 121
Figure 20 Italian terra sigillata in Pannonia:map 123
xv
Figures
Figure 2I Inscription set up by the burial-club of the cives Agrippinenses 125
Fig/ire 22 Legionaryfortresses:canabaeand municipium 127
Figure 23 Canabaelegionis: Carnuntumand Aquincum 128
Figure 24 Mines and municipia in Upper Moesia 132
Figure 25 Ulpii and Aelii in the territorium of Aquincum and Ulpianum 146
Figure 26 Monumentsto nativesin Pannonia 149
Figure 27 Auxiliary diplomata from Pannonia:map 15 6
Figure 28 Detailed plan of the municipium of Aquincum 160
Figure 29 Town plans 2: Sirmium, Carnuntum 163
Figure 30 Detailed plan of the municipium of Carnuntum 165
Figure ]I Town plans 3: Aquincum, Bassiana,Scarbantia,etc. 167
Figure }2 Villa plans I: Parndorf,Donnerskirchen,Eisenstadt,etc. 170
Figure 33 Villa plans 2: Bataca,Poganytelek,Smarje 17 2
Figure 34 Plans of native houses 174
Figure 35 Inscription of Commodusmentioninglatrunculi from
Intercisa 196
Figure 36 Distribution of coins of Regalianusand Dryantilla 207
Figure 37 The progressof urbanization:map 220
Figure 38 Distribution of tombstonesin Pannonia 233
Figure 39 Stationsof benejiciarii 235
Figure 40 Mithraea in Pannonia 255
Figure 4I Plans of templesand shrines 257
Figure 42 Fortified landing-places 270
Figure 43 The 'Devil's Dyke' 271
Figure 44 The late Romanprovinces:Pannonia 274
Figure 45 The late Romanprovinces:Upper Moesia 275
Figure 46 Late Romanforts in the Djerdap 281
Figure 47 Plansof late Romanforts in Pannonia 283
Figure 48 Late Romanforts on the Danubebend 292
Figure 49 Plans of late Romanvillas and houses 301
Figure 50 Late Romanfortified settlements 304
Figure 5I Distribution of Christianinscriptions 309
Figure 52 Plans of Christian cemeteries 314
Figure 53 Christian basilicae: funerary monumentsin Sopianae 31 6
Figure 54 Plan of the vicus of the auxiliary fort at Matrica 318
Figure 55 Christian tombstoneof Aurelius Iodorus, a Greek from Lao-
dicea, found at Savaria(HungarianNational Museum) 33 1
Figure 56 Christiantombstoneof Artemidora buried near St Synerotas
at Sirmium 333
xvi
Figures
Figure 57 Paintedvault of the cella memoriaeat Pecs(Sopianae) 334
Figure 58 Geographicalnamesof ancientorigin 355
Two folding mapsat end:
Before Plates
Figure 59 Map of Pannoniashowing placesmentionedin the text
After Plates
Figure 60 Map of Upper Moesia showing placesmentionedin the text
XVll
Preface
There are various ways in which the history of a Roman province may be
written. In preparingthis book I was not much attractedto a routine descrip-
tion, subject by subject, of the history of events or wars, administration,
economiclife, of town and country, arts, religion and so on. One of the chief
problems in the study of the history of the Roman empire is that research
relating to the empire as a whole is not organically interconnectedwith that
relating to its parts. A daunting gap separatesthe study of central Romanim-
perial history from local, often highly developed,archaeologicalresearch.This
gap may be bridged only by the use of a method which exploresevery aspect
period by period and in accordancewith historical principles. This method I
havechosendespitethe fact that the resultsof local researchat presentavailable
arenot as uniformly useful as might be desired.
The readerwill certainly notice that I have not always beenable to apply my
principles equally in both the provinces assignedto me. It would have been
ideal for my purposehad researchbeen conductedin Pannoniaand in Upper
Moesia with identical aims and with equal completeness.But such studiesare
still decisively influencedby the history of the last hundredyears, and even in
Pannoniaitself the desireduniformity hasnot yet beenachieved.Even now, the
estimableefforts of severalgenerationsof Austrian, Hungarian,Jugoslavand
Czechoslovakscholarshave scarcelyachievedthe adoptionof generallyagreed
methodsand principles of study and field-work.
Nevertheless,many attemptsat individual synthesisare available: I was able
to basemy studieson much previous work of great value. If in my footnotes
and in the selectbibliography I have failed to give them the prominencethey
xix
Preface
deserve,the omission is not due to ingratitude but to the universal fate of
scholarly studies,which this book in its turn will not escape.
In my treatmentof Upper Moesia I have had to face fundamentaldifficulties
and lack of information. In spite of the encouraginggrowth of Serbianresearch
today, thereis still so much work to be undertakenthat I have beencompelled
to generalizefrom all too few data, work how I would. Nor must it be over-
looked that, owing to the very fact that Upper Moesia is only half the size of
Pannonia,less spacecould here be devotedto the former; moreover,for much
of the periodits Danubefrontier was of minor political and military importance.
A further decisionof methodhad to be made: were the two provincesto be
treated separatelyin two parts under one cover, or parallel and togetherin a
comparativestudy as a subdivision of Illyricum? I have chosen the second
method,for it seemedto me that the history of the two provincesis complemen-
tary. Evenif, on the larger stageof the Romanempire,they fail to play the part
of an intelligible unit of history, in their own area they enjoy many close
connections.
I had also to solve the problem where to sever the threadsconnectingthe
history of our provinceswith that of the empireas a whole. Someof my readers
will no doubt feel that I have cut them off too short, while otherswill miss the
amplification of local detail which might be expectedto form the strengthof a
good provincial monograph. But it would be pretentiousto claim that the
presentvolume is a referencebook: its more modestaim is to offer a synthesis
of what I myselfhavelearnt-andin part developed.It mustbe left to the reader
to decide whether the synthesisoffered is premature.I believe, however, that
any seriousstudentof Romanprovincial archaeologymust sooneror later write
his monograph.
This work, then, is no authoritative source-book,but the product of an
archaeologist'surge to synthesize.It is necessarilysubjective, since aspects
which I believe to be important are emphasized,while others are left in the
backgroundor neglected.I haveendeavouredto give due coverageto political,
social and economicconditions, and to the history of civilization and religion.
In the footnotesI have given primary sourcespriority over modernliterature;
where sourcesare numerousI refer to authorswho have listed and evaluated
them. Neither the footnotes nor the select bibliography contain referencesto
standardworks on Romanimperial history; theseare readily availablein more
generalworks. In compiling the select bibliography I have kept in mind the
needsof scholarsby including publicationsin which referenceswill be found to
the older but still usefulliterature.
My original text was written in Germanin 1970. Later information if pub-
xx
Preface
lished by the summerof 1972 is includedin the bibliography; it appearsin the
text only if I have beenpersuadedto modify my original views. But someim-
portant recent discoveries could not be given the extendedtreatment they
deserved: for instance, the reconstruction of the Scarbantia street-system
(K. Sz. Poczy), the newly excavatedRoman building in barbarianterritory in
Slovakia (T. Kolnik), the excavationsof late Romanfortified settlementsin the
interior of Pannonia(K. Sagi, S. Soproni, E. Toth), the large-scaleexcavations
at Sirmium, the recentlypublisheddiscoveriesat municipium Dardanorum(the
late E. CerSkov),andespeciallythe vast programmeof Jugoslavexcavationsin
the Iron Gates(the resultsof which, however,arenot yet fully available).
I must thank T. Nagy, who allowed me to make use of his important dis-
coveries concerning the topography of the legionary fortress of Aquincum
(Figs 14,22 and 23). It is a particularpleasureto have beenable to refer in more
than one context to work of my own pupils, and to have been compelledto
reconsidermy own opinionsin consequence. Indeedthe views expressedin this
book are the outcomeof twelve yearsof lecturingin the University of Budapest
and of the stimulatingcontactwith my pupils which hasresultedfrom it.
I am indebtedto severalfriends for readingthe whole or parts of my manu-
script, and am particularly gratefulfor help given by my colleaguesL. Balla and
J. Fitz. Above all lowe much to the friendly criticisms of J. Gy. Szilagyi, which
enabledme to clarify variouspassagesin the text and to avoid someerrors.Any
mistakesand obscuritiesthat remain are my own responsibility.
I wish to expressmy particular thanks to ProfessorSheppardFrere for his
careful supervisionand correctionof the English translationof my text and for
mucheditorial work in the preparationof this volume.
Budapest A. M6csy
XX!
Chapter I
Thracians, Illyrians and Celts
At the time when the Greekswere alreadyfamiliar to someextentwith the geo-

graphyand ethnographyof central Europethe upperDanubewas firmly in the
handsof the Celts, whereasthe lower reaches,from the Iron Gatesto the Black
Sea,hadfrom time immemorialbelongedto the settlementareaof the Thracians
and the Getae. While, of course, some details concerning the races who
inhabited the valleys of the western and southerntributaries of the Danube
betweenVienna and the Iron Gates were available to the Greeks, they were
scarcelyadequateas a basis for an authoritative ethnic appraisal.The earliest
information is to be found in the Iliad: 'Zeus ... turned his shining eyesaway
into the distance,where he saw the lands of the horse-rearingThraciansand
the M ysoi who fight hand to hand... ! I Hellenic philologists queried this
passagein Homer, as the Greeksin that era knew only of the existenceof the
M ysoi in M ysia, part of Asia Minor. It was Posidonius,the last great scholarof
Hellenism, who provided the right answerby pointing out that a tribe called
the M ysoi lived north of the Thraciansand could be identified as the M ysoi
mentionedby Homer.2
This old controversyis to some extentcharacteristicof the whole of the
Greeks' knowledgeof their northern neighbours.In archaic times they had a
pretty clear picture of the Danube river-systemand of the inhabitantsof the
Adriatic coast. This information, gathered by Greek traders, was in part
forgotten in the classicalperiod, along with the decline of Greektradewith the
north, and in part becameintermingled with mythology and hencegradually
distorted.3 It became,for example, a self-evident truth that one arm of the
Danubeflowed into the Adriatic and hencethat the Balkanswere an island.4It
I
Thracians, Il!yrians and Celts
was not until Roman expansionto the north that a fresh breezebearing new
knowledgeput an endto what was in the main idle speculationand information
derived from books. Even as far as the Mysoi-Moisoi were concerned,Posi-
donius' information was only scanty, whereasa Roman army under the com-
mand of C. ScriboniusCurio, advancingfrom Macedoniaas far as the Danube
between76 and 72 B.C., broughtback more up-to-datedetails of the inhabitants
of the Danubevalley.s Greek scholarswere interestedin the history and ethno-
graphy of central Europe only in so far as the barbarianswere involved in
events in the Greek homeland. This episodic information which has been
handeddown is, however,adequatefor a sketchto be madeof the main features
of political developmenton the middle Danubein the last centuriesB.C.
Apart from theM ysoi referredto by Homer,the earliestpeopleson the middle
Danubementionedby nameare thoselisted by Herodotusand Hecataeus,e.g.
the Sindoi, Sigynnoi, Kaulikoi, etc., none of whom playedany part in the sub-
sequenthistory of this area.6 Later they were not included amongthe peoples
who lived in the Danubevalley.7 They were, however,tribes who were still in
contactwith Greektrade,whereaslatertribesfor the mostpart only camewithin
the Greek and Romanrangeof vision if they had becomethe enemiesor allies
of Macedoniaor Rome. It has recently been assumedthat the Sigynnoi, who
probably lived in the great Hungarianplain, traded with the Veneti in Istria,
who in turn were involved in early Greek trade on the Adriatic coast.8 Traces
of thesetrading contactsare attestedby a few late archaicfinds in Hungaryand
the northernBalkans.9 In the fifth centuryB.C. the ethnographicpicture at least
of present-daySerbiabecomessomewhatclearer(Fig. I). According to Hero-
dotusthe valley of the Morava(which he calls the Brongos)below Nis belonged
to the Triballi, a tribe centrednorth of the Balkan mountainson the Danube
aroundOescus(Bulgarian: Gigen), andthesepeoplelater often provedtrouble-
some to the Macedonians.The western neighboursof the Triballi were the
Illyrians, for the Angros, a tributary of the Brongoswhich cannotbe identified
with certainty(Ibar? ZapadnaMorava?Toplica?), rises in Illyrian country and
flows into the Brongos in the area of the Triballi. So much for Herodotus.IO
Later sourcesII report the advanceof the Illyrian Autariataeinto the territory
of the Triballi about the end of the fifth century B.C. and not much later there
are reports of Celtic conquestsin the Carpathianregion and on the lower
Danube.After the Autariataehad driven the Triballi out of the Morava valley
they themselveswere subduedat the beginning of the fourth century by the
Celts,12who were emergingas a new influencein the Carpathianregion and in
the Balkans.The known ethnic patternof the original inhabitantsin the Roman
period first beganto take shapeas the result of the Celticization of many areas
2
Thracians, Ilfyrians and Celts
kilometres
e
n ub
Da
le
m iddb e
the u
t hDe a n
id dle
m
he the
t h et
e
be
be
ub
u
u
n
an
an
Da
D
D
le
le
le
d
dd
id
id
i
m
em
em
e
nub
th
th
th Da
em iddle
iddt the m
lehe th
Dm em
a nid th idd
udble em le
e Da idd Da
nu le nu
be Da be
nu
be
land over 1500

kilometres
? 20? kilometres
Pigure I The areaof the middle Danubein the fifth century B.C.
3
in south-eastEurope,andit is only after the arrival of the Celts that it is possible
to follow the political changeson the middle Danubemore precisely.
The Triballi belongedto the Thraco-Geticethnic group which inhabitedthe
easternhalf of the Balkanpeninsula.The Autariataewerean Illyrian tribe which
had settledin the westernhalf. The third ethnic and linguistic component,the
Celts, probablycamein part from northernItaly and in part along the Danube
from the west.I3 The Thracians,Illyrians and Celts were the threemost impor-
tant ethnicandlinguistic groupsin south-eastEurope,andin Romantimes they
madeup the nativeinhabitantsof PannoniaandMoesiaSuperior.The linguistic
boundaries ran through these two provinces: the Celtic-Illyrian through
Pannoniaand the Illyrian-Thracian through Moesia Superior.14 This is one
reasonwhy sizeablepolitical units were only rarely established.Neither Pan-
nonia nor Moesia Superiorwere known as geographicor political conceptsin
pre-Romantimes; both regionsbelongedfor the mostpart to political structures
which had their centresoutsidethe country.
As far as the ThraciansandIllyrians are concerned,the problemssurrounding
their origin and linguistic classificationare today much in a stateof flux. There
was a tendencyin the last decadeto reject the theory of linguistic uniformity
among the Thraciansas well as among the Illyrians, and to reservethe terms
'Thracian'and 'Illyrian' for a smaller and more readily definabletribal group.
After the abandonmentof suspectPan-Illyrism in the 1930s,philologists came
to recognizethat the Veneti and Liburni on the northernAdriatic coast spoke
a languagerelatedto, thoughdifferent from, Illyrian. 15 Analysis of the earliest
information about the Illyrians on the Adriatic indicatesthat the nameIllyrioi
applies only to a small areain the south of what was later to becomethe pro-
vince of Dalmatia;16 a critical classificationof Illyrian names17 has established
that there were two or three distinct areasin the provinces of Pannoniaand
Dalmatia.Finally, doubtswere also raisedby archaeologistswho attributedthe
late BronzeAge and early Iron Age culture of northernDalmatia to a people
who differed from the Illyrians.18 The Dalmatiansand Pannonians,therefore,
were either not Illyrians or at bestwere only relatedlinguistically; but they did
not speakthe samelanguageas their neighboursto the south, who alone are
regardedin the sourcesas Illyrians. In the caseof the Thraciansthe situationis
even more confused. Classical scholars were convinced that the Getae and
Daciansspokethe samelanguagel9 and that the Thraciansand Getaewere in
fact one and the same people.20 The mapping of the place-namesin Dacia,
Moesia and Thracerecently producedthe hypothesisthat the Thraciansin the
wider sensebelongedto two different linguistic groups: partly to the Thraco-
Getic and partly to the Dacian-Moesian. 2I This division is basedprincipally on
4
the suffixes (-dava, -para, -sara,etc.), the geographicaldistribution of which is
to some extent restricted. Historically there wouldprobably have been better
justification for a division into a Dacian-Getic group and a Thraco-Mysic
group; argumentbasedon and restrictedto place-namesis too narrow a basis
for suchsweepinghypotheses.In any casePannoniaand Moesia Superiorwere
on the peripheryof both the Illyrians and the Thraciansin the wider senseof
those terms. For the moment it will perhapsnot be misleadingif, following
tradition, thesefringe racesare regardedas Illyrians and Thracians,though we
must not overlook the fact that dialects-probablymore so then than now-
affected linguistic uniformity. A number of tribes are known to have existed
along the linguistic boundaries,right inside Pannoniaand Moesia Superior,
thoughthe linguistic and ethnic groups to which they belongedhave not been
conclusivelyestablished.Of the tribes whosehistory will be describedlater, the
Dardaniare variously takento be of Illyrian and of Thracianstock;22 the Scor-
disci were, of course,a group establishedby the Celts; in imperial times, how-
ever, their nameswere Illyrian (Pannonian),23and some sourcesinclude them
amongthe Thracians;24 it is only recently that the Eravisci have beenconclu-
sively identified as a Celtic race.25 The contactzonebetweenIllyrians, Thracians
and Celts was obviously a very broadone, and their relationswith one another
were subjectto constantfluctuation, which ceasedonly with the Roman con-
quest.
Celtic expansionreachedthe Carpathianarearoughly at the sametime as the
Celtic invasion of Italy, that is at the beginning of the fourth century B.C.
According to Celtic legend,300,000peoplemigratedto Italy and Illyria: Livy's
accountmentionsBellovesusand Sigovesus,nephewsof the Celtic king Ambi-
gatus, who sent them againstItaly and againstthe inhabitantsof the Herrynia
silva respectively.26The Celtic legend in Pompeius Trogus describes wars
against the native inhabitants which lasted for years and led to the gradual
subjugationof the Pannonians. 27 Early La Tene finds in Pannoniasuggestthat
the Celts advancedalong the Danubeand conqueredonly the north-westpart
of the Carpathianregion in the fourth century B.C.28 (Fig. 2).
Another arm of Celtic migration to the eastprobably startedfrom northern
Italy, for, while fighting on the Danube, Alexander the Great received an
embassyfrom the 'Adriatic Celts'.29 This incident coincided with the violent
collapseof the Autariataewho, accordingto somesources,were forced to yield
to the Celtic advance,and after a long and turbulent wanderinghereand there
were finally wiped out.30 Towardsthe end of the fourth centurythe Celts, who
had already establishedthemselvesin Pannonia,renewedtheir raids on the
Balkan peninsula; these soon brought them into conflict with the king of
5
Danub e e
Danub
u be
e u be Dan
b a n
beanu D
uD
ea n
ubD
e an
n ub D
Da
Da
nu
Dba
en
u be
Dan
uDbae
nub
e h
fo urth e fourt
i n trhthe th
beou
D anntuhe f Da
i
ube nu
Dba
Dan en
u be
Danub
Da Danub e
nu ein the in the fourDth an
Da fourth ub
beD nu ein
inan be Da the
thub in nu fou
e fei Da th b rth
oun t nu e
e f in
rthhe be ou t
fo in r t hhe
ur th fo
th ef ur
ou th
rth
land over 7500

metres --~
o- 200
, Kilometres
Figure 2 The areaof the middle Danubein the fourth century B.C.
6
Thracians, Il!Jrians and Celts
Macedoniaon the northernborderof Thrace.31 Theseraids were the preludeto
the great Celtic invasion of the Balkans in 279 B.C., which, in many respects,
resembledthe migration of more than a century earlier. On this occasion,too,
the numbersinvolved seemto have beenvery large; Justin in his epitome of
PompeiusTrogus mentions a chargeby 150,000infantrymen under Brennus'
command,32and even if this figure is grossly exaggerated,since Diodorus
mentionsonly 50,000,33this muchis certain:the enterprisewas concernedfrom
the outsetwith finding new areasin which to settle. Again Celtic legendindi-
catesthat there were two leaders,Belgius (or Bolgius) and Brennus,whereasa
variant statesthat there was a third group under the commandof Cerethrius.34
A sectionof the peopleand of the troops stayedbehindto protectthe homeland
(oikeia),35 which can probably be identified as the area in Pannoniaalready
consolidatedby the Celts. From the dating of a Greek bronzevesselfound in a
gravein the La Tenecemeteryat Szob on the Danubebendin Hungary(Pl. 1a)
the only conclusionis that it was booty from Greece.36
The years 280-278 were markedby fluctuating battleswhich are of no con-
cern here.37 Brennus' army plunderedDelphi; one of his bands which had
desertedcrossedinto Asia Minor, while a third group was able to establishan
independentCeltic statesouth of the Balkan rangewhich persistedfor several
decades,'the empireof Tylis'. Brennus'army, split into severalbands,withdrew
to the north in 278. It is probablethat thesegreat migrations, which created
more or less permanentCeltic territories in the Balkan peninsula, brought
aboutan intensiveCelticizationof the areassurroundingthosefrom which they
had started,namelyin the Danubevalley from below Vienna to the Iron Gates,
where in fact there had beenearlier Celticization (Fig. 3). Tradition has it that
the establishmentof a very strongpolitical structurein the north of present-day
Serbia can be traced back to thesemigrations: the Celts on their return from
Delphi settled on the mouth of the Save and called themselvesScordisci.38
Moreover,it is probablytrue that thoseparts of Pannoniawhich had not been
conqueredin the first Celtic invasion were alreadyCelticized by the beginning
of the third centuryB.C. La Tenefinds of the C periodaredistributedthroughout
the whole of what was later Pannonia,although it is impossible to decide
whether those in certain areas, e.g. in the mid-Save valley, indicate Celtic
settlementor merely the distribution of artefactsmadeby Celts. The areaunder
the control of the Scordiscihas also yielded fairly rich finds of La Tenetype.39
The races north of Macedoniadid not immediately recover from the up-
heavalsof the Celtic invasion. Sourcesmentionvarious rearguardbattlesin the
Balkans;40 heavybattlesprobablyalso resultedfrom the consolidationof Celtic
control in the Danubevalley. From the middle of the third century onwardsa
7
kilometres
kilometres
s
s e tre
e tsrielom
eksi loemtrek
m eektrsilom
sil e r
o t
e trielkom
lom k
kil
ki
om
kilo
etr
met
res
eksilo
m
kilo
etr
kilo
me met kilo
es
trke res me
islom trke
etr islom
eksi etr
lom eksi
etr lom ki
es etr lo
e s m
et
re
s
kilo kilo
me me
trke trke
islom islom
etr etr
eksi eksi
lom lom
etr etr
es es
land over 7500

metres----,I
o
,
200
I kilometres
Figure} The areaof the middle Danubein the third century B,C,
8
Thracians, Il(yrians and Celts
very active power was emergingin the immediateneighbourhoodof the Mace-
donians;andbeing probablythe first to free itself from the Celts, it was expand-
ing to the east, south and west. This power consistedof the Dardanianrace
which had settledon the upper reachesofAxius (Vardar) and in the Kosovo
polje and Metohija basins north of the huge ScardusMons (Sar Planina). It
belongedto the Thraco-Mysic Balkan group41 which was racially connected
with the peoplesof westernAsia Minor (Mysoi in Asia Minor, Moisoi on the
Danube,Dardanoion the Axius and near Troy, Phrygiansin Asia Minor and
Brigoi in Thrace,etc.).42
In the fifth century the Dardanianswere probably subject to the Triballi.
Whetherthey were reachedby the expansionof the Autariataeis unknown,but
it is not improbable.In the time of Philip II they fought against the Mace-
donians43 andthey probablysufferedseverelyat the handsof the Celtic invaders
at the beginning of the third century. In 279 B.C. Brennus'tribe was in Dar-
dania,44and it was therethat Leonnoriusand Lutarius (who transferred20,000
Celtsto Asia Minor) desertedhim. Celtic rule in Dardaniadid not, however,last
long. About the middle of the third century the Dardanianswere already
revealingthemselvesas very dangerousenemiesof the Illyrians, and by the last
third of that century they had becomethe proverbial enemiesof Macedonia.45
The Macedoniankings, Demetrius, Antigonus Doson and Philip V had to
wage constantwar againstthem. The battle-groundswere in the Vardar gap
around Stobi and south of the ScardusMons, and the wars were basically
defensiveagainsta mountainpeoplewhich was constantlyengagedin raiding.46
Philip V wantedto apply tough measuresin order to put an end to this raiding
onceandfor all, but beforehe could do so the Dardanianraids suddenlyceased
in 197. This is all the more surprisingsince in the sameyear the Macedonians
were defeatedat Cynoscephalae,so that the cessationof the Dardanianattacks
cannotbe attributedto an increasein strengthof the Macedonianforces.More-
over the Dardanianswere increasingly beingregardedas the natural allies of
Rome, and thus would have beenin a strongposition for launchingsuccessful
attacksagainstMacedonia.
It is very probablethat from the beginning of the secondcentury the Dar-
danianswere being threatenedfrom the north and were concentratingtheir
forces there. Thereis also probablya connectionwith the fact that antagonism
betweenRome and Macedoniahad drawn the Balkan peoplesinto strategic
alliances.For sometime after 278 B.C. nothingis heardof the Scordisciwho had
establishedCeltic control to the north of the Dardanians.Then, under Philip
V, they suddenlyemergeas allies of the Macedonianking and as the enemies
of Romeandthe Dardanians.Whereasnothingis known of what was happening
9
on the Dardanians'northernboundaryor on the Scordisci'ssouthernboundary
in the early decadesof the secondcentury, the events of the year 179 throw
unexpectedlyclear light on the pattern of political alliances, and from this
developmentbetween197 and 179 may be inferred.
In 179, on the basisof an earlieralliance,Philip V was instrumentalin making
the Bastarnae,possibly a Germanic (or Celtic?) tribe on the mouth of the
Danube,set out to conquerDardaniaand, having doneso, to move on to attack
Italy. This attack was to be launchedfrom the north, and with this in view
Philip V had alreadyformed an alliance with the Scordisciwho controlledthe
route to Italy. This large-scaleplan failed, as in the war (which lasted several
years)the Dardanianssucceededin putting the Bastarnaeto flight in 174.47 The
years 179-174were catastrophicfor the Dardaniansdespitetheir victory over
the Bastarnae,and resultedin their being soundly defeatedin 170 by Perseus,
the last Macedonianking.48 From these events it is possibleto pick out the
salient aspectsof political conditions prevailing in the central Balkan areaand
to the north of it; in the secondhalf of the third centurythe Scordisciconquered
the Savevalley, of specialimportanceas the only route to Italy, and as a result
gradually becamethe most important power in the northern Balkans. The
Dardanians,who shortly after 278 were able to free themselvesfrom Celtic
domination,threatenedMacedonia'snorthern boundaryand thereforebecame
potential allies of Rome, whereasthe Scordisciemergedas the natural allies of
Macedonia.As a result, the Dardaniansfound themselvesbetweenthe devil and
the deep sea, and despite their successfulstruggle for freedom against the
Bastarnaewere unableto hold out againstthe Scordisciand the Macedonians.
The disturbancein the balanceof power in the central andnorthernBalkansin
the early decadesof the secondcentury was the preludeto almost a century of
hegemonyby the Scordisci.They becamenot only successors to the Dardanians
on the northern borderof Macedonia,which was alreadyunderRomancontrol,
but time and again fought againsttheir westernand south-easternneighbours
and, it would seem,subjugatedDardanians,Moesians,Pannonians,as well as
severalother tribes.
There is no information available concerningevents between279 and 179
in the areawhich was later to becomethe provinceof Pannonia.The establish-
ment of Celtic hegemonyunder the Scordisci presupposesthat Celtic control
north of the Saveand south and west of the Danubehad been consolidated,
and that it provided adequatebacking for their supremacyin the south of
Pannonia.The gradual Celticization of Pannoniaprobably took place in the
third century (Fig. 3); clear effects of this are to be found in native namesin
Romantimes. This Celticization was so thoroughin the northernhalf and on
10
::/ . d an Celts
. s Ilhrtans
Thraczan,
kilometres
kilometres
kilometres
kilometres
kilom
kilometres
I
etre
C
kilom
s
kilo
me
tre
etre
s A
kilometres
s
kilome
kilometres D
tre
kislome
treksilo ki
melotrm
esetre
kislom etreks ki
ilomet lo
res m
et
re
s
kilo
me
kil trke
om islo
et m etr
rke es
islo kilometres
m kilometres
et kilometres
rek kilo
silo
m me
et tre
re ksilo
s me
tre
s
land over 7500
metres----,I
o,
200
I kilometres
. the second centuryB.C.

Figure 4 The areaof the ill1.ddle Danube 1ll
II
the western edge of Pannoniathat the pre-Celtic languagecompletely disap-
peared.It was lesseffective southof the river Drave-asin the caseof the Scor-
disci themselveswhosenamesin Romantimes are Illyrian (Pannonian).49This
is obviously due in part to the numericalratio of Celtic conquerorsto the non-
Celtic native populationand in part to the length of time Celtic control lasted.
The Celtic group, which after its retreatfrom the Balkansin 278 B.C. was able
to establishitself under the name Scordiscion the mouth of the Save,formed
only a thin Celtic upper class,which was graduallyabsorbedby the subjugated
but neverthelessnumerically stronger native population. It is therefore not
surprising that this people is described jn the sources sometimesas Celts,
and sometimesas Thracians,while by the end of the secondcentury B.C. they
are actually given an Illyrian suffix and are called Scordistae.so Accordingly the
Scordisci,who were frequently the only, and always the strongestpower in the
centralBalkanareain the secondcentury,areto be regardedas merelyoneof the
Celts' political creationsand not as a Celtic tribe.
After Rome conqueredMacedonia the Dardanians were heard of once
againwhen they laid claim to PaeoniaYThus by 168 the Scordiscihad not yet
extendedtheir control to include Dardania.Around the middle of the second
century it is probablethat their interestlay rather to the west, for in 156 they
were defeatedby the Romans,S2 possibly in the great Dalmatianwar in which
they fought on the side of Rome'senemies.The first battle on the Macedonian
borderbetweenthe Romansand the Scordisciis not attesteduntil 141; S3 there
are no records of earlier battles againstMacedonia'snorthern neighbours.In
156, probably during the Dalmatian war, the Romans laid siege to the Pan-
noniantown of Siscia(Segestike)at the mouth of the Kulpa in the Savevalley S4
andthis advanceto the eastprobablyalso involved the Scordisci.It is aboutthis
time that the first historical mention of the Pannoniansoccursin a fragment of
Polybius.ss
It is possiblethat it was this defeatin the west, which probablyoccurredin the
upper Savevalley or in northernBosnia,that causedthe Scordiscito turn their
expansionsouthwards.In the middle of the secondcentury they must have
subduedthe Dardanians,since from 141 onwards only the Scordisci together
with west Thracian tribes are constantly mentionedas being the enemieson
Macedonia'snorthernboundary(Fig. 4); someof the battlestook place where
the Macedoniankings had earlier engagedthe Dardanians.For a long time after
the middle of the secondcenturythereis no further mentionof the Dardanians.
The permanentallies of the Scordisciin their many wars between141 and 109
were the west Thracian tribes, among whom the Maidoi are repeatedlymen-
tioned; the theatre of operationsincluded the valleys of the Axius (Vardar),
12
Astibus (Bregalnica),Strymon (Struma) and the upper reachesof the Hebrus
(Marica).56 It must therefore be assumedthat the area controlled by the
Scordisciin the secondhalf of the secondcentury coveredthe whole of what
was later to become Moesia Superior. According to Strabo, this control
extendedas far as Paeonia,Illyria and Thrace,and involved the subjugationof
many tribes in the central Balkans.57 Strabo also mentions that the Scordisci
living west of the Morava were known as 'the big ones'and those to the east
as 'the little ones'.58
While their hegemonylasted,the Scordiscidid not give up their claim to the
Savevalley. When in 119 B.C. the RomansbesiegedSisciafor the secondtime,59
a Dalmatianwar was in progressin which the Scordisciwere again involved.
And when someyearslater the Cimbri were migrating throughthe westernhalf
of the Carpathianregion, onceagainit is only the Scordisciwho are mentioned
as living in the Savevalley.
The migration of the Cimbri did not bring about any substantialchangesin
the history of the Danubelands,but it is neverthelessof particularinterestsince
Posidonius'descriptioncontains many important details, for instancethat the
Cimbri first cameup againstthe Boii and, having beenrepulsedby them, met
with the Scordisci, the Taurisci and then the Helvetii.60 This list of names
provides a clear picture of the power structurein the Carpathianregion at the
end of the secondcentury. The Cimbri migratedfrom the north through the
westernhalf of the Carpathianregion, the northernpart of which was controlled
by the Celtic Boii, the southernpart by the Celtic Scordisci and the south-
westernpart by the Celtic Taurisci. It is not known where the boundarybe-
tween the Boii and the Scordisciran, but it is very probablethat the border
betweenthe Taurisci and the Scordiscilay west of Siscia. This can be deduced
not only from the fact that the Scordiscitook part in the battles eachtime the
RomansbesiegedSiscia, in 156 and 119 B.C., but also from a passagein Strabo
which statesthat the upper reachesof the Save belongedto the Taurisci, but
Siscia to the Pannonians. 61
As for the Pannonians,Posidoniusdoesnot even mentionthem. Thus at the
end of the secondcentury they must have been controlled by the Scordisci;
Polybius, however, writing of an event which took place not much earlier,
indicates that he was aware of their existence.62 The Pannonians belonged to
the Illyrian or to the pre-Celtic native populationwhich was relatedlinguistic-
ally to the Illyrians and which inhabitedthe north-westernBalkan area;accord-
ing to detaileddescriptionsby Straboand Appian, they broke up into a number
of tribes which settledin a fairly big areaextendingfrom the Drave as far as the
Dardaniansand the Ardiaei in southernDalmatia.63 Only a few of thesePan-
13
nonian tribes belonged to the region which later became the province of
Pannonia;of the tribes in the province of Pannonia,Strabo classesonly the
Breuci and the Andizetesas Pannonian.In Romanimperial times the Andizetes
were living at the mouth of the Drave and had the Breuci as their southern
neighbourson the Save.In addition to thesetwo tribes, othersliving between
the two rivers must also be regardedas Pannonian,namely those called Cola-
piani, lasi, Oseriates,Amantini, Cornacatesand Scordisci, all within Roman
Pannonia.Names appearingon inscriptions from these tribal areasform the
northernsectionof the so-calledcentral Dalmatian-Pannonian group of names
which hasits distributionin the areaoccupiedby the Pannonians,i.e. in southern
Pannoniaand easternDalmatia.64 The original inhabitantsof the north-western
part of what was later Moesia Superiorwere possibly Pannoniantoo, but at an
early stage they had come under the control of the Scordisci, whose centre
happenedto be in that very area.
In Posidonius'descriptionof the migration of the Cimbri there occurs the
first mentionof the control of north Pannoniaby the Boii. As alreadyindicated,
this areawas underCeltic occupationfrom the beginningof the fourth century,
and not only the wealth of La Tenematerialbut also the obviously Celtic names
of the native inhabitantsof western and north-westernPannoniain imperial
times point to an intensive Celticization. Namesof tribes in the fourth, third
and secondcenturieshave, however, not been handeddown. Posidoniuswas
the first to reveal that control was in the handsof a tribe which had the same
name, Boli, as a once powerful Celtic tribe in northernItaly. Tribes with the
same name frequently occur in different countries, either becausethey are
racially connected(subdivisionoccurring only later as the result of migration),
or becausethe tribal name in the languageconcernedwas a common ethnic
term, or becauseduring its migrations the tribe appearedat different placesat
different stagesof history. The easiestexplanationis, of course, to assumea
migration, theconclusionwhich Posidoniusreachedwhen he suggesteda racial
link betweenthe north Italian Boii and the Boii in the Carpathianarea.Accord-
ing to this hypothesis,the Boii were driven out of northernItaly at the begin-
ning of the secondcenturyand went to the Danube,65or, accordingto another
ancientversion,to Bohemia,the ancientnameof which, Boiohaemum,provided
welcomesupportfor this theory.66Whetherthereis any historical basisfor this
idea is opento question.Thoroughinvestigationof La Tene material doesnot
indicate so close a relationshipto finds in Bohemia-partof the original Celtic
homeland-thata Celtic migration from Bohemia into Slovakia may be de-
duced.67 There is even less evidenceof any link with Celtic finds in northern
Italy.68 All the more striking is the close relation betweenthe Slovakian and
14
HungarianLa Tenematerial,from which we may infer a gradualCelticizationof
the whole of the northern Carpathianregion between the fourth and the
secondcenturies.69 This process-perhaps without further Celtic immigration
from the west-startedwith the first Celtic invasion at the beginning of the
fourth century; if the Celts in the north-westof the Carpathianregionwere in a
positionat the beginning of the third century to embarkon further large-scale
migrationsinto the Balkansand evenas far as Galatia,then it must be assumed
that gradualCelticizationwithin the Carpathianareaalso occurred.
Power-relationships,as indicatedby Posidoniusand Strabo,soonunderwent
a radical change(Fig. 5). In the first place it may be assumedthat there was a
loosening of Scordiscancontrol in Dardania. In 97 B.C. the Dardanians,of
whom nothing had beenheardfor a long time, appearedon the sceneas the
allies of the west ThracianMaidoi, and were defeatedby Macedonia'sRoman
army.70Not much later the Dardaniansand Scordisciwere allied in the struggle
againstSulla on Macedonia'snorthern boundary,7Iand shortly afterwardsthe
Scordisciwere defeatedby L. Cornelius Scipio Asiagenus,a defeatfrom which
they never recovered.It is impossibleto date Scipio's war preciselywithin the
years 88-81 B.C., as the relevant sourcesgive rise to much uncertainty; for
instancethe war is said to havebeenstartedin revengefor a secondplundering
of Delphi by the Celts. There is also a greatlack of clarity in the chronological
order of events.7ZNevertheless,one thing may be assumedas certain, namely
that Scipio gaineda truly decisivevictory. The Scordiscinever againappeared
as dangerousenemieson Macedonia'snorthernboundary,and between81 and
15 B.C. there is only one further mention of them. According to Appian they
withdrew to the mouth of the Saveand to the Danubeislands;this is, of course,
a gross exaggeration.In fact they disappearedonce and for all from the area
which was later known as Moesia and thenceforth are mentioned only as
inhabiting the south-eastcorner of Pannonia.
It would not be in keepingwith availableevidenceif the declineof Scordiscan
power were attributed solely to Scipio's military victories. Macedonianand
Roman campaignsagainst barbarianson the northernMacedonianboundary
always had a limited aim, the establishmentof peace,and there was never any
decisiveinterferencein the power-relationshipsof the tribes in question.The
Scordisciwereprobablyunderpressurefrom other directions,and eventswhich
occurredsoon afterwardsmake this a very reasonableassumption.When, for
example,in 64 B.C. Mithridatesset out to attackItaly via the northernBalkans-
a plan which the last Macedoniankings had also seriouslyconsidered-itis the
Pannoniansand not the Scordisciwho were saidto control that area.73 Thus the
Pannonianswho, apparently,had liberatedthemselvesfrom the dominationof
15
kilom
etres
I
I
O
B
kilom
etreksilo
metre
kilom
etres
s
kilometrekislo
metres
kilometrekislo
kil
metres k
ilo
om
me
etr
ttrree
kksisillo
eksil
om me kil
ettrre kil om
o
om
me
ess et
et re
tre
re s
s
s
kilo
me
kil trke
om islo
et m etr
rke es
islo kilo
m
et me
rek tre
silo ksilo
m m
et etr
re es
s
land over 7500

metres----,I
o 200
, I kilometres
Figure J The areaof the middle Danubein the first half of the first century B.C.
the Scordisciin the early decadesof the first centurywere regardedas being the
dominantpowerin the Savevalley. This roughly coincidedwith the strengthen-
ing of the Dacian state, now united under Burebista, and he, in turn, soon
attackedand defeatedthe Scordisci. As a result the latter becamethe allies of
the Daciansin their strugglesagainstthe Celts in the Carpathianarea.74
So far it has not been possible to provide conclusive evidencefor all the
changesin power-relationshipsand the ethnicpicturein the first half of the first
centuryB.C. (Fig. 5). For instance,it has only recentlybeenacceptedas probable
that the migration southwardsof someCeltic tribes, describedby Caesarin his
Commentariesas a Helvetian migration,75 involved the Carpathianregion. The
Latobici in the upper Savevalley probablyarrived then, whilst the nameof the
Hercuniatesin eastPannoniaperhapssuggeststhat they also originatedin the
north. The Latobici obviously brought with them from the Saale area their
strangecustomof shapingtheir cremationvesselslike houses.76 A questionstill
awaiting clarification is that of a westernmigration by a sectionof the Boii who,
accordingto Caesar,besiegedNoreia, the capital of Noricum.77
Theseconvulsionsin the central Balkansfinally led to a particularismwhich
producedmany small statescomposedof tribes and tribal groups which acted
individually (Fig. 5); the place of the Scordisci was taken by the Dardanians,
Moesiansand Pannonians,all of them tribes which in the secondcentury had
depended,closely or loosely, on the Scordisci.
At first the Dardaniansprovedtroublesometo the Romans.In 76 B.C., along
with Thracian tribes, they invadedMacedonia,and the war which ensued(at
first under the leadershipof Appius Claudius Pulcher,78then of C. Scribonius
Curio and finally of M. TerentiusVarro Lucullus) lastedfrom 76 to 73. This
bellum Dardanicum,79which Curio wagedenergeticallybut with unprecedented
cruelty, rangedover a very wide area.80 It was at this time, too, that a Roman
army reachedthe Danubefor the first time, an event whose significancewas
often underlined by later Roman historians.81 The Roman army probably
advancedthrough the Isker (Oescus)valley and brought news of the tribe of
the Moesi, which was settledin the Timok valley and on the Danubebut had
not beenheardof before.82 It is highly probablethat it was this newswhich led
Posidoniusto interpretthe passagein Iliad xiii, 5, correctly. Homer vouchesfor
the fact that the Moesians were long-establishedinhabitants of the eastern
Balkans. But for many centuries nothing was heard of them, as their more
powerful neighbours,the Triballi, Autariatae, Dardaniansand Scordisci had
either held them in subjectionor pushedthem into the background.It was only
after the power of the Scordiscihad declinedthat they emergedas an indepen-
dent political force.
Curio also madecontactwith the Dacians,but this did not lead to a head-on
collision betweenthem and the Romans.The Scordisci are not mentionedin
connectionwith this campaignfor, after their defeatby Scipio, the questionof
whetherthey were enemiesor allies of Rome doesnot arise. At the time of the
campaignsof 76-73 the Scordiscihad not yet beendefeatedby the Dacians,and
as the latter are not mentionedin 64 when Mithridates wanted to send the
Bastarnaethroughthe Savevalley into Italy,83 their defeatof the Scordiscimust
have occurred later. The approximate date of the Dacians' victory can be
establishedby meansof the following considerations.It resultedin their tem-
porary supremacyin the centralBalkans.84 In the last yearsof Caesar'sdictator-
ship Romeregardeda Dacianattackon Macedoniaas not unlikely, and after his
murder there were in fact rumours that such an invasion had taken place.85
Thusfor a time the Dardanians,too, musthavebeenunderDaciancontrol. They
were, however,still independentwhen in 62 and 57 they were attackingMace-
donia.86 Therefore, the Dacian victory over the Scordisci occurred roughly
between56 and 50 B.C.
According to Strabo, Burebistacarried out his conquestswithin a matter
of a few years.87 It is not necessaryhereto go into detailsregardinghis successes
in the south-east.His operationshad, however, a lasting effect on the history
of the Carpathianregion, inasmuchas they put an end once and for all to
Celticizationin many areas.
As has alreadybeenpointed out, political control in the westernhalf of the
Carpathianbasinand in the valley of the Saveat the turn of the secondcentury
was still in the handsof the Celtic Boii, the Scordisci(who werein the last resort
Celts) and the Celtic Taurisci. In the Save valley the Pannonianssoon made
themselvesindependent;in the north, however, in the first half of the first
century,the BoH were still in uninterruptedcontrol. The easternneighboursof
the Boii and the Scordisciwere the Dacians;Caesar,who was familiar to some
extentwith the ethnic and political conditionsin the Danubevalley (whetheror
not he derived his information from Hellenistic geographyand in particular
from Posidonius),knew that the easternend of the enormousundefinedareaof
the Her0'nia silva to which the northernCarpathiansbelongedwas inhabitedby
the Daciansand the Anartii, and that thesetwo tribes were to be found eastof
the Danubebend.88 Ptolemy also mentions that the Anartii lived in northern
Dacia,89 andhe is also awareof the existenceof the Taurisci, the latter'sneigh-
bours.Thesetwo Celtic tribes obviously belongedto the Boian tribal federation
which embracedthe whole of the northernhalf of the Carpathianbasin. The
late Iron Age oppidumculture can be tracedfrom the Danubebendas far as the
north-eastern Carpathians (Budapest-Gellerthegy,Zemplin, Mukacevo, to
18
mention only the most important centres). Further tribes representativeof
Celtic control are mentionedin Tacitus'Germania,andthese,togetherwith later
information, make it possibleto draw up a list of virtually all the tribes on the
northernedgeof the Hungarianplain: the Osi on the Danubebend, and to the
eastthe Cotini, Anartii and the Taurisci.90 The zone of contact betweenCelts
and Dacians must, therefore, have run roughly through the middle of the
Carpathianregion. Strabo,who gives a fairly detaileddescriptionof the battles
betweenthe Celts and Dacians, statesthat the boundaryline was the Parisos,
which is the ancientnamefor the river Tisza.91 Accordingto Strabo,Burebista's
claim to certainareasled to war which endedc. 45 B.C. in the defeatof the Boii
and their allies the Taurisci. Burebista'sopponentwas Critasirus,the king of
what was probably a very large country. The Taurisci, over whom he ruled,
were Celts in south-westPannonia,and the Boii, who were also Celts, lived in
northernPannonia;to them the Celtic tribes on the northernedgeof the plain
and in Slovakia belonged.Within the Romanempire Burebista'svictory made
a deepimpression.Therewas referenceto a 'Boian desert'in Pannonia forwhich
Pliny the Elder usesthe term deserta.92
The Boii were no more wiped out than were the Scordisciby Scipio. But the
extensiveand apparentlywell-establishedBoian area of control disintegrated,
while the Dacians,evenif they did not found any large settlements,nevertheless
establishedoutpostsin manypartsof the Carpathianregion. Typical small hand-
made dishes with handles have been found in the later layers of late Celtic
oppida: theseDacian dishes and various other Dacian ceramicshave beenun-
earthedin the Carpathianregion, particularly in thosepartsinto which the Boii
and the Scordiscihad extendedtheir control, e.g. the Banat,the Morava valley,
in east and west Slovakia and also here and there on the right bank of the
Danubein the north-eastof what was later to becomePannonia.The numberof
locationsis particularly striking in Slovakia where Dacianizationcontinuedin
imperial times.93 After the Cotini had beensettledin Pannoniaand Moesia by
the EmperorMarcus Aurelius some of them were mentionedon inscriptions;
in the main they haveThraco-Daciannames.94 Later on, Dacianswho had been
driven out of the Tisza areaby the Iazygesalso settledon the northernedgeof
the plain.95
Daciancontrol of the westernhalf of the Carpathianregion did not, however,
last long. Burebistadied about44 B.C. andhis statebrokeup into at leastfour or
five kingdoms.96 The tendencytowards particularism, inherent in barbarian
political structures,again put an end to the plans of a great ruler. The new
Dacian kings, Burebista'ssuccessors,were, of course,involved in the struggle
for power betweenthe Roman parties after the murder of Caesar,97but the
19
kilometres
kilomektrileometr
s es
tres
s k ilome
s re
trmeet kilometres
e
es
o m
k ilo
il
esetr
I
k
kielotrm
kilom
C
ki ilom
ki
lo
k
lo
m etreksilo
m
et
et
re
rek
s
A
silo
kilo
m
kilometres me
et
trke
me
re
islo k
kilometres me ilom kil
s
ktriel
tre etreom
osm
s kisloetr
mekeislo
ki
D
ekitlro
lo
tre me m
ki
s tre
em
lo
s ke
et
s
m
rek
itlroe
et
silo
re
ms e
s
et
kilo
tre
re
me
s
s
kil trke
om islo
et m etr
rke es
islo kilo
m
et me
rek ktrielo
silo kislm
m ometr
et ekterisl
re eosm
s etr
es
land over 7500

metres----,I
o 200
, I kilometres
Figure 6 The areaof the middle Danubein the secondhalf of the first centuryB.C.
20
greatking's Dacian stategraduallycontractedand soonlost its leadingposition.
In 39 the Dardanians,who had again become independent,invaded Mace-
donia;98 it is possiblethat the Pannoniansreconqueredsome areasbelonging
to the Celts,and the Daciansdisappearedonce and for all from what was later
to becomePannonia.They were able to maintain their hold only on the plain,
where later-it is not known when-they had to yield the area betweenthe
Danube and the Tisza to the Iazyges. In his description of Dacia, Ptolemy
gives the Tisza as its westernboundary,99and so it remaineduntil Dacia was
conqueredby Trajan.
The only political power which was not defeatedby any external enemy
was that of the Pannonians. This group of tribes, in so far as its capacityto form
a statewas concerned,was perhapsthe weakest.It is of coursea commonplace
to mention anarchyin ethnographicdescriptionsof barbariantribes. Appian,
however, in his referenceto the Pannonians,does not confine himself to the
merereiterationof suchplatitudes:'The Pannoniansdo not live in towns, but in
villages and hamlets organizedon the basis of clans. They do not assemblein
joint councils,nor do they have joint leaderswho are supreme;100,000of them
are capableof bearing arms, but becauseof the prevailing anarchythey never
assembleas a combinedforce.'100 This primitive tribal societyprovidesadequate
explanationof why the Pannoniansdid not appearon the stageuntil after their
better-organizedneighbourshad wiped each other out. After Caesar'smurder
there was no power in the Carpathianregion nor in the central Balkans which
could seriously have opposedRome. Nor were the wars waged by Roman
generalsin these areas after c. 44 B.C. dictated by the necessityto intervene
radically in the inter-tribal relations obtaining in the Danubearea: both Octa-
vian's campaignagainstthe Iapodes(35-33 B.C.) and M. Licinius Crassus'war
againstthe Dacians,Moesiansand Thracianscan be understoodonly against
the backgroundof the political situationin Rome itself.
In Caesar'slast years Burebistawas in undisputedcontrol of the Carpathian
region and of the north-easternpart of the Balkan peninsula.Caesarhad the
rumour circulatedthat he was formulating a large-scaleplan for crushingthe
Dacianking,101 and Octavian,Caesar'sexecutor,had to fit a Dacianwar into his
political plans,102despitethe collapseof Dacian hegemonyin the meantime.In
this contextit was rumouredthat certainDaciankings would sidewith Octavian
or with Antony; the rivals accusedeachother of having formed an alliancewith
the Dacians.103Even after Actium the 'Dacian problem' remaineda topic of
conversationin Rome,I04
It is only againstthe backgroundof this Dacian problem, inflated for pro-
paganda purposes,that Octavian's campaign against the Iapodes becomes
21
at all intelligible. In 35 B.C. he advancedagainst the Iapodes, the Alpine
peopleto the eastof Aquileia, on the pretext that they had ceasedpaying their
taxes.IOSAfter hard-foughtbattleshe succeededin capturingtheir most impor-
tant fortresses,including Metulum. From here he continuedhis advanceinto
Pannoniancountry 'althoughthey had given no cause'.106The final goal of this
advanceseemsto havebeenthe captureof Siscia, twice unsuccessfullybesieged
by the Romans(in 156 and againin II9 B.C.). This strong-pointon the mouth
of the Kulpa in the Savevalley was the natural spring-boardfor an advanceto
the east,and it was in this context that Octavian'spropagandaput the capture
of Siscia. It was said to be the most important basefrom which to launch an
attackagainstthe Dacians.Appian, in his accountof the waragainstthe Iapodes,
which is indirectly based on Octavian'scommentaries,reports a quarrel be-
tween the upperclassesandthe commonpeopleof Siscia. The former were in
favour of yielding to the Romanterms-thehandingover of 100 hostagesand
acceptinga Romangarrison: thepeople,however,resistedand in the end the
town hadto be takenin a battlewhich lasteda month,I°7This accountby Appian
is at variancewith his above-mentioneddescriptionof the primitive social con-
ditions among the Pannonians.The suspicion is justified that Octavian, to
defend his attack, was reverting to the trick repeatedlyused by Caesarin his
Gallic war, the suggestionthat a sectionin the enemycamp-alwaysthe aristo-
cracy-wasfriendly to Rome, and that, therefore, the Roman general came
not as a conqueror but as the supporter of the Romanophile aristocracy.
Such reasoningin support of the capture of Siscia was all the more neces-
sary becausethe Pannonianswere not the traditional and proverbial enemies
of Rome. Even Appian rightly admits that over a long period the Romans
had taken no notice of the peoplesliving on the other side of the Eastern
Alps,I°8
After capturingSiscia, Octaviandivided the town into two by meansof a wall,
and occupiedit with a force consistingof twenty-five cohortsunder the com-
mand of Fufius Geminus. He made the Pannoniansmake submissionto him
without continuing his advanceeastwards,and then returnedto Rome for the
winter. In the following year there is again no mention of an advanceagainst
the Dacians:in 34-33 Octavianpacifiedthe Dalmatiantribes living to the south
of the Iapodes.
It is now clear that Octavian'splans did not include a Dacian war after the
captureof Siscia. The three-yearwar undertakenon this pretext109 had ulti-
mately resultedin the pacification of an areaof great importance.By securing
the EasternAlps and the coastalstrip along the northernAdriatic, a link was
establishedwhich might be of importance not only in an advanceagainst
22
Antony but alsofor the future occupationof Illyricum, which hadbeenawarded
to Octavianat Brundisiumin 40 B.C.
Thereare no reportsof Siscia'sfate in the following decades,apartfrom the
fact that a rising by its inhabitantsin the winter of 35-34B.C. was put down. It
is howeverprobablethat in practicaltermsit remainedin Romanhands.In the
wars underAugustusit becamethe Romanarmy's most importantstronghold.
Had not its fortressbeenfirmly held by the Romans,Tiberius' campaignagainst
the Scordisciin 15 B.C. would have beenimpossible.
The Romanadvancefrom Macedoniawas likewise influencedby the Dacian
problem. The generalwho set aboutputting Caesar'splan for a war againstthe
Daciansinto effect cameinto conflict with Octavianfor that very reason,andhis
victories were eliminatedfrom the official versionsof the history of that war,II°
M. Licinius Crassus,a former supporterof Antony, who went over to the side
of Octavianshortly before Actium, receivedthe consulship,along with Octa-
vian, in 30 B.C. (without having previously held the praetorship:he probably
insistedon the consulshipas a rewardfor changingsides).In 29 he becamepro-
consul of Macedoniaand in this capacityhe launcheda large-scalewar on his
own account.
This war, in which Crassuswas victorious againstthe Thraciansand Getae,III
was officially known as the Thraco-Geticwar. However,it may be inferredfrom
a brief note by CassiusDio that the first peoplesto be defeated,and the real
enemies,were the Dacians and their allies, the Bastarnae,II2Horace extols
Crassus'war in the words occidit Daci Cotisonis agmen;II3 the later accountis,
however,influencedby the official version,accordingto which Crassusdefeated
only the Moesiansand various west Thraciantribes. The Dacians,defeatedby
Crassus,werethe subjectsof King Cotiso,whoserule also extendedsouthof the
Danube.II4 He had taken over the territory in the central Balkans which had
been conqueredby the Dadans.In view of the way history was distortedfor
political reasons,it is difficult to reconstructthe first part of this Dacian war,
even in outline; nevertheless,the additional uncensoreddetails provided by
CassiusDioIIS make it possibleto come to some conclusionsabout it. After
mentioningbriefly the victory over the Dacianshe prefaceshis descriptionof it
by the remark that the Dardanians,Triballi and the Dentheletaewere defeated
by the Bastarnae,andthat Crassuscameto the aid of the Dentheletae.Now it is
known that the Bastarnaewerea tribe of mercenariesliving at the mouth of the
Danube,who frequently put their troops at the disposalof the Macedonians,
Mithridates, the Dacians,etc. It is probablethat at this time they were taking
part in the war on the side of the Dacians,and it is readily understandable why
they in particularshouldattackthesetribes. The latter were theneighboursand
23
natural enemies of the Dacians who were in control south of the Danube.
Cotiso, when attackedby Crassus,probably appealedto the Bastarnaefor help
and they then attackedthosetribes which wantedto get rid of Dacian control
in the neighbouringarea, and were consequentlypotential allies of Rome. It
was the Dentheletaewho askedCrassusfor help, and after defeatingthe Dacians
he advancedinto Moesianterritory. In the following year (28 B.C.) he continued
his war in westernThrace.
The war conductedby Crassusnot only broke Dacian control south of the
Danubebut also led to the pacification of the Dardanians,Triballi, Moesians
and some of the west Thraciantribes. From 28 B.C. onwards there was peace
in the central Balkans.Then in the year 16 there were reportsof an invasion of
Macedoniaby the Scordisci and the Dentheletae. I16 In the sameyear the Pan-
noniansinvadedIstria. CassiusDio, who recordsboth theseinvasions,saysthat

the Pannonians'allies were theNoricans;thesewereprobablyCeltsliving in the
upper valley of the Save, that is to say they were the Pannonians'western
neighbours.CassiusDio goes on to say that the repulseof the Pannoniansby
P. Silius Nerva led to their renewedsubjection; the indication that there had
beenan earlier one probablyrefers to Octavian'scampaignagainstthe Iapodes.
It is very probablethat it was theseinvasionswhich madeAugustussendthe
young Tiberius, probably as early as 15 B.C., against the Scordisci and some
Thracian tribes, neighboursof Macedonia. This campaignmight also be re-
gardedas a step towardsthe pacificationof the borderregions of north-eastern
Italy and northern Macedonia,II7A consequenceof Crassus'Dacian war was
that the Scordiscihad again achievedindependenceand, occupyingas they did
a key position at the mouth of the Save,could representa dangerfor both Italy
and Macedonia.However, when, a few years later, Tiberius had to fight the
Pannonians,he was backedby allies in the Save valley who were none other
than the Scordisci.I I 8
Further eventsin the westernCarpathianregion and in the central Balkans
took place as a result of the Romanadvancetowardsthe Danubefrontier, and
belong, therefore,to a new chapterin the history of that area(p. 34).
Literary sourceshave made it possibleto trace the history of the last four or
five centuriesB.C., at leastin its main features.As for the social structureand the
basis on which statestructuresoften developedrapidly to embracewide areas,
the sourcesareby no meansso helpful. Nevertheless,from the previousaccount
certain inferencesmay be drawn which can easily be reconciledwith the very
sparsedirect information containedin ancientliterature.II 9
One of the most striking characteristicsof all the political structureswas the
24
dominant position of one tribe within each of them usually extendingover a
wide area.At the beginningof the periodfor which thereareliterary sourcesthe
Triballi exercisedsupremeauthority over a very considerableregion of the
centralBalkans.Their own settlementareaand basewas a not particularly large
territory lying betweenthe Balkan mountains(Haemus)and the Danube,where
their existenceis attestedright into imperial times.I2O Their centreat that time
was at Oescus(Gigen). They were driven out of their western and southern
possessionsby the Illyrian tribe of the Autariatae,who supersededthe Triballi
in the fourth century until they were themselvesdriven out by the Celtic
onslaught;after succumbingto this they disappearedalmost entirely from the
historical scene.The Celts had establisheda similar, though more permanent,
political structurein the westernpart of the Carpathians.In the courseof the
third centurythey wereableto extendthe areaundertheir control with astonish-
ing speed.At the sametime the Dardaniansestablishedsupremacyin the south,
although it did not extend over so wide an area as that of the Triballi and
Autariatae. In the second century B.C. it was known that there were Celtic
tribes (Boii, Taurisci and Scordisci) exercisingindependentcontrol within the
area dominatedby the Celts. The Daciansthen made their appearanceunder
King Burebistaandestablishedbrief control over a variety of non-Daciantribes.
Thesepolitical structureswere eachcentredupon a tribe which had cometo
the fore as conquerorand organizerand upon various local tribes, not neces-
sarily related to it linguistically; the local population was either subduedand
exploited in the harshestmannerpossible,or was forced into an alliance. For
eachof theseextremeforms of treatmentthere is clear evidence.At the height
of their power, accordingto Theopompus,the Autariataehad 300,000subjects
'in the conditionof helots';121 aboutthe middle of the first century,on the other
hand, the Scordisciwere forced to take part in Burebista'swars as the allies of
the Dacians.I22 The military expeditionsconductedby the Triballi, Autariatae
and the Celts were, therefore,probablystartedby a martial sectionof the tribal
society. This enterprisingand mobile warrior classwas able from time to time
to conquerlarge areasand to exploit their inhabitants.There is probably an
historical basis for the legend surroundingthe Celtic princes Bellovesusand
Sigovesus,123not to mention Brennus,Bolgius, Leonnorius,Lutarius, Cereth-
rius, Akichorius and others who set out with their wives and children around
the year 279 and subsequentlysucceededin establishingCeltic states in the
Balkansand evenin Asia Minor.
It is less easy to give a clear definition of the circumstancesand conditions
which determinedthe duration and effect of thesepolitical formations. At the
beginningof the fourth century, Celtic bandshad conqueredthe north-western
25
part of the Carpathianregion and set aboutits Celticizationwith determination.
During the empirenon-Celticname-elements in north-westPannoniaarealmost
completelylacking. Celtic bandsalso establishedcontrol through the Scordisci,
though later sourcesrefer to the latter as a mixed tribe composedof Celtic,
Illyrian and Thracianelements,while the namesfound in imperial times in that
part of south-eastPannoniaoccupiedby the Scordisciclearly belong to the so-
called centralDalmatian-Pannonian group and revealno Celtic influenceworth
mentioning. Celtic bands also set up the 'empire of Tylis' in Thrace, but all
trace of them disappearedin the third century.124In the absenceof archaeolo-
gical and linguistic evidenceit has not beenpossibleso far to establishthe loca-
tion of this Celtic state. At the same time as it came into being, the Celtic
kingdom of Galatia was also being founded; this becamean all too dangerous
rival of Pergamumand was able to survive.
The impact of Celtic influence on Pannonianmaterial culture was successful
even in areaswhere there was evidencein Roman times that the indigenous
population werenot Celts. But there was no such fundamentalchangein the
latter'sburial rites. Inhumation,as practisedby the Celts,nevergainedthe upper
hand, although such burials here and there in early imperial times go back to
Celtic elementsin the original population.I 2S Becausethe late Iron Age in Pan-
nonia has not beenadequatelyinvestigated,and in Moesia Superior even less
so, it is impossibleat presentto decidewhethertherewere considerablediffer-
encesin the spreadof La Tene culture which could be attributedto variations
in the degreeof Celticization.In any caseLa Tenetypesare generallycharacter-
istic of local productionin pre-RomanPannonia.Recentexcavationshavepro-
ducedLa Tenefinds in the northernpart ofMoesiaSuperior,in the territory of
the Scordisci;126 in the southernpart of this province,wherethe Scordisciwere
only temporarily in control, the influence of La Tene culture is not attested.It
was this area which, from the middle of the fourth century at the latest, was
able to preserveits independenceon a permanentbasis. The Dardaniansdid,
of course,suffer severelyduring the Celtic invasion of 279, during that of the
Bastarnaein 179 and probably ultimately at the hands of the Scordisci but,
retaining their old name, they maintainedtheir identity until imperial times.
It is also possibleto say somethingaboutthe socialstructureof this tribe. Greek
ethnographershad a certainamountof information about the Dardanianswho,
as the neighboursof the Macedonians,had come intocontactwith the Greeks
at an early date. Hence,thereare frequentreportsof undergroundhuts covered
with dung, of their proverbialdirtiness,of musicalinstrumentsand last but not
least of productssuch as cheeseand woollen goods.I 27 All this suggestsa race
of mountain shepherds.Agatharchidessupplies the further information that
26
Thracians, Il(yrians and Celts
therewas a rich Dardaniansocial classwhosememberskept thousandsof serfs
who had to till the land and do military servicefor their masters.12S It is pro-
bable that here there was a symbiosis of warrior mountain shepherdsand
peasantskept in a state of subjection. This social structure was doubtless
similar in many respectsto that in which one tribe becamedominant,and hence
was in a position to repulseattacksby externaltribes.
The Dardaniansbelongedto the older group of Balkan peopleswho were
racially relatedto thoseof north-westAsia Minor. This relationshiphadalready
struck the Greeks: 'There are many similar names among the Trojans and
Thracians,' writes Strabo.I29 The Trojan Dardanians,attested in the Iliad,
promptedwriters in late antiquity to representthe emperorswho camefrom
Moesia Superior, Constantinethe Great and Justinian in particular, as being
the descendantsof a Trojan, i.e. an ancientRoman, race. The Moesi (Mysoi),
who first appearin the first centuryB.C. after the defeatof the Triballi, Autaria-
tae, Scordisciand Dacianswho followed one anotherin the control of Moesian
country, also belongedto the Thraco-Phrygiangroup of races.What the pre-
Celtic native inhabitants of Pannoniaand the north-westernpart of Moesia
Superior were called is not known, the reasonbeing that Celtic control over
theseareasled to fundamentalethnic and social changes.The inhabitantsof the
areasover which the Scordisciestablishedcontrol were later known as Scord-
isci, althoughtheir languagewas not Celticized-onthe contrary,they absorbed
the Celtic ruling class. The only pre-Celtic section of the population in Pan-
nonia whose name is known is the Pannonians,but it is not known how far
their original territory extendednorthwards.Like the Moesiansthey did not
appearon the sceneuntil after the break-upof the hegemoniesestablishedby
individual tribes. At that time, in the first half of the first century,they lived in
the northernpart of what was later Dalmatia, in the Save valleydown-stream
from Sisciaand on the lower reachesof the Drave.I3O It is very possiblethat they
were the original inhabitantsof that part of Pannonialying betweenthe Drave
and the Danube,I3I but there their languageunderwentcompleteCelticization.
What information we have concerningPannoniansociety we owe entirely
to Appian, whosebrief descriptionhas alreadybeenquoted.In contrastto the
Celts and Daciansthe Pannonianswere unable, even temporarily, to establish
any kind of political unity. Their state of anarchy, underlined by Appian,
showeditself evenin their resistanceto Augustus,when eachtribe fought as a
separateforce underits own leadersagainstthe Romans.Hencethey were not
capableof submitting even temporarily to a central power on the lines of that
establishedby Critasirus the Boian, or Burebistathe Dacian. Their primitive
social institutions possiblyexplain why nothing was heardof them until a late
27
date in the Danube and Balkan areas. The better-organizedand technically
superior Celts subduedthem without difficulty; it was not until the Celtic
Scordiscihad been defeatedthat they appearedas an independenttribe in the
Savevalley.
The degreeto which the Celts were superiorto the nativepopulationcan best
be understoodby consideringthe circulation of the so-calledbarbariancoin-
ageI32 (Fig. 7). Money as an economic adjunct was not used in the central
Balkansor in the westernhalf of the Carpathianareauntil after the consolidation
of Boian and Scordiscanpower. All the coins are copies of various Greek
tetradrachms,in particular the Philippus, though these did not themselves
circulate to any marked extent either in Moesia Superior or in Pannonia.I33
Thus the copieswere not a replacementfor money already in circulation, but
representedthe first coinageto serve a function in local economiclife (Pl. 2).
The introduction of moneyoriginatedwith the Celtic tribes and evenlater was
generallyrestrictedto those areasunder Celtic control. Mapping of coin-finds
shows concentrations,in the region of Sirmium, attributableto the Scordisci
(Fig. 7); in the Danubevalley from Vienna to the Sirmium region, whereCeltic
control was alreadyfirmly establishedin the fourth to third centuries;and in a
strip stretching from the Danube as far as the Carpathians,in the north of
present-dayHungary and in Slovakia, which can be traced backto the Celtic
advancein the third to secondcenturies. That south-westPannoniahas not
producedvery many coin-finds to date is perhapsdue to the fact that Roman
money beganto circulate there at an early stage(Fig. 7).
It is impossibleto stateprecisely just when theseprimitive coins first began
to be minted. The standardauthority on Celtic coins in the Carpathianarea
dates the earliest mintings to the end of the secondcentury;134 more recent
Slovakianspecialistsfavour a century earlier.13sIn actualfact the period of the
Celtic invasionof the Balkans,whenthe Celts first cameinto contactwith Greek
money, must be takenas the earliestpossibledate.Moreover, Celtic coins were
first found in excavationsat late Celtic oppida along with moulds for dies and
occasionallytools for striking the coins.136It was preciselyin the late Celtic
period that tradeflourished; certaintypesof bronzewares,weaponsand jewelry
spreadover a remarkablylarge area,indicating a very uniform oppidumculture.
Both Polybius and Posidoniusknew that there was vigorous trade along the
Save. The entrep6t was Aquileia where, according to Strabo, the barbarians
from the Danubearea exchangedtheir goods, slaves,cattle and skins against
wine, oil and productsof the sea.137Thesecommoditiescannot, of course,be
attestedarchaeoiogically.Thereis, however,evidencethat Italian families in the
latter part of the republican period owned slaves with Illyrian names138 and
28
an
Rom
Ro
ma
n
Ro
ma
n
an
Rom
Ro
m
RoRoam
n
ma an
n
• Celtic
o 100
,
200
Greek & Roman I
( Hoards only) /(f/ometres
Figure 7 The distribution of Greek, Romanrepublicanand native coins

Thracians, II!Jrians and Celts
also that Italian merchantsin the sameperiod establisheda trans-Alpinebranch
in Nauportusnot far from Ljubljana.I39 Romanmoneydid not, however,reach
Pannoniaand Moesia Superior until after the conquestof the Save valley by
Octavian140 (Fig. 7). Money obviously servedonly the needsof local trade,and
consequentlythe area in which the individual types of Celtic coins circulated
was fairly small. Most coin-finds representa narrowly restrictedseriesof related
types. In contrastto those of other barbariancountries,the coins do not bear
chieftains' names, probably becausethey were minted not by the chieftains
themselvesbut by traders whothen had them circulated within the radius of
their activity.I4I
The only exceptionis the region of Sirmium, whererepublicandenarii, Greek
tetradrachmsand coins of Apollonia and Dyrrhachiumare frequently found,I42
andwhere,accordingto K. Pink, the minting of coins by the easternCelts first
started(Fig. 7). The importantpart playedby the Scordisciin the history of the
Balkan peninsula makes the circulation of coins readily understandable,or,
more correctly, the important role of the Scordisci may be attributed largely
to the fact that they continuedto occupy an areawhich was the meeting-point
of the most important routesin south-eastEurope. Romancoins and thoseof
the Adriatic coastaltowns reachedthe Sirmium region along a route which,
startingfrom the Adriatic, crossedBosniain a north-easterlydirection. This link
explains not only why the Scordiscitook part in the Dalmatian wars but also
why under Tiberius this very route was developedas one of the first roads in
the Balkans.The reasonwhy the Scordisciwere crushedby Tiberius in 15 B.C.
also becomesclear at last: their territory was of the utmost importancefor the
control of both Pannoniaand Moesia.
Chapter 2
Conquest of the Danube region and
Augustan frontier policy
Rome'sinterestin the region eastof the Alps was first arousedwhen minerals,
primarily silver and iron, beganto be mined in the areasoccupiedby the Celts.
Rome'sfirst and most important approachto theseareaswas the founding of
Aquileia, at the mouth of the Natiso, in 181 B.C., after fluctuating battleswith
the Celtic tribes of the north-easternAlps. The establishmentof this harbour
was of vital importancein the subsequenthistory of the Danubeprovinces.
Therethen followed a variety of attemptsto advancenorthwards,culminating
ultimately in the successfulexploitation of Norican iron on Rome's behalf.
During the second century Aquileia had become an extremely important
Mediterraneanentrep6t;here, iron from Noricum was loadedon to ships; and
from here the trade of north-eastItaly and that of the north-easternAlps was
organized.The extentto which the merchanthousesin Aquileia were involved
in the exploitationof Noricanmineralsis attestedby recentAustrianexcavations
on the MagdalensbergnearKlagenfurt.
East of the Julian Alps there was no subsidiarydepot from Aquileia similar
to that on the Magdalensberg.WhereasAppian's I/!Jrike containsa great deal
of information concerning the early history of Illyrian-Roman relations, it
particularlyemphasizesthat theIllyrian racesbeyondthe Alps werelong ignored
by the Romans.I The mannerin which the Romansgradually obtainedcontrol
of Norican iron, without at the sametime formally conqueringand occupying
Noricum, indicatesthat they were not concernedwith the annexationof Nori-
cum as a provincebut solely with acquiringcontrol over its mineralsand trade.
Iron andsilver meantmoreto the Romansthanthe slaves,cattleandskins which
the Illyrians sold in the market at Aquileia. Romantraderswere concernedto
3I
Conquestof the Danuberegion and Augustanfrontier policy
keep control over Norican iron on the mining sites north of the Alps, whereas
the barbariansin the Savevalley were interestedin transportingtheir goodsto
Aquileia in order to sell them there.2 Rome consequentlymade no diplomatic
efforts to involve the racesof the north-westBalkansand of the Savevalley in
Mediterraneantrade; on the other handshefrequently had to wagewar against
the Alpine tribes, particularly the Taurisci and the Iapodes,who, as neighbours
of Aquileia, either threatenedthe town itself or were even more of a threat to
the latter's expanding vineyards and other important agricultural estates.
Furthermore,the inhabitantsof the Julian Alps turned the transport of goods
over the Alps betweenthe Save and Aquileia to their advantageby acting as
guides, delaying or hijacking it (in the ancient literature they are describedas
tribes of brigands),3 so that from time to time the Romanswere forced to take
rigorous action againstthem. As long as the Macedoniankingdom lasted, an
advanceby that power into the Savevalley was fearedin Rome,as it could open
up the route to Italy. An advanceon a larger scalebeyondthe Alps into the Save
valley took placeonly when the Liburnian and Dalmatianpiratesthreatenedthe
interestsof Aquileia, that is to say its shippingon the Adriatic. In 156 and again
in II9 B.C., on both occasionsin the context of Dalmatian wars, the Romans
laid siegeto Siscia. However,military operationshaving the expresspurposeof
occupying the upper Save valley were first carried out by Octavian. The
campaignconductedby the consul C. CassiusLonginus in 171 seemsonly to
havebeenan attemptto establisha link betweennorthernItaly andthe Balkans.
On the pretext of attackingMacedoniahe advancedover the Julian Alps into
the country of the Istri, Iapodesand Carni, but returnedafter plunderingtheir
lands.4 Thesetribes, allies of Rome, complainedbefore the Senatein 170. It is
possible that CassiusLonginus only intended to intimidate Aquileia's neigh-
bours. Otherwisethereis no evidencethat Rome was bent on making strategic
use of the anti-Macedonianattitude of some Balkan races.
The Romans were even less interested in advancing northwards from
Macedonia. The bold campaignsof Scipio Asiagenus, ScriboniusCurio and
Licinius Crassushad only the limited aim of pacifying the northernpart of the
province of Macedonia,which, as in the days of the Macedoniankingdom, was
subjectto invasion by various Balkan tribes.
As already mentionedin Chapter I, accountsof the campaignsof Octavian
(35-33 B.C.) and Licinius Crassus(29-28 B.C.) were distorted for propaganda
purposesto suggestthat both were preventiveactionsagainstdangerfrom the
Dacians.In fact, after the deathof Burebista,who ruled over a centralizedstate
and controlledlarge areasacquiredby conquest,the Dacian kings ceasedto be
dangerousenemies.Octavian'saim in advancingtowardsthe Savevalley must
32
Conquestoj the Danuberegion and Augustanjrontier policy
consequentlybe tracedto the needto establisha land-link betweenItaly andthe
Balkans and at the sametime to pacify troublesomeneighboursof the north-
easternItalian colonies, primarily Aquileia and Tergeste.That these Alpine
tribes were successfullypacified is attestedby two late Republicaninscriptions
from Nauportus(Vrhnika, south of Ljubljana), which indicate that there was
a settlementof representativesof Aquileian trading houseson the other side of
the Alps.s In any caseit was only underAugustusthat Romandenarii first began
to circulatein what was later Pannonia,thoughthey did so before I I B.C., which
is regardedas the official date of the annexationof Pannonia.6 In 35 B.C. the
Pannoniansin the Savevalley madesubmissionto the Romans.7 This prompted
Octavianin his speechto the army beforethe battle of Actium to point out that
Romansoldiershad advancedas far as the Danube.8
Not much later Crassusalso advancedfrom the south as far as the Danube.
This campaignwas the last Roman military operationthat we know of in the
areawhich later becameUpperMoesia.Thereis just as little evidencethat formal
annexationof this country was envisagedas there was in connection with
Octavian'sactivities in Pannonia.Whereasthe latter was formally incorporated
into the empire in I I B.C. and the imperial frontier pushedforward to the
Danube,the province of Moesia was not establisheduntil much later and then,
so to speak, unobtrusively. The problem of Moesia's beginningsis still un-
resolved; it has, however, much in common with that of the annexationof
Pannonia,whose larger northernarea was only occupiedgradually and in the
sameunobtrusivemanner.
The questionwhen and how Moesia was occupiedand becamea province
has puzzled modern historians, since from Crassus(29/28 B.C.) to Poppaeus
Sabinus,9the first governor, there was an interval of roughly half a century
during which there were neither wars of conquestnor rebellions. For a long
time it remainedan open questionwhether Pannonia'snorthern boundaryin
Augustan times was the Drave or whether that emperor advancedto the
Danube,as claimedin the MonumentumAncyranum,lofor the Pannonianwarswere
waged in the Save valley and there is scarcelyany evidencethat the Danube
frontier was occupiedin Julio-Claudiantimes. Thus, heretoo, thereis a gap of
half a centurywhich neitherarchaeologicalmaterialnor historical datacan fill.
EventsunderAugustusin the northernBalkansandin the westernCarpathian
regions shed light on these related questionsand at least make it possibleto
formulatethe problemitself more preciselyand henceto seeit in perspective.It
will becomeclear that wars of conquestand the reductionof rebellionswere no
more preconditionsfor the establishmentof a province than was a permanent
systemof garrison-postsand fortressesthroughoutthe new province.
33
In the previous chapterthe purposeof the campaignwaged by the young
Tiberius againstthe Scordisci (15 B.C.?) was interpretedas being to pacify a
peopleholding a key positionand to turn them into allies of Rome.The pattern
for this operationprobably derived from Romanexperiencesubsequentto the
captureof Siscia.To beginwith, it provedby no meanseasyto pacify the Celtic
and Pannonian(Illyrian) inhabitantsof the Savevalley. In fact, in the winter of
35/34, immediately after Siscia was taken, they rebelled, and in 16 B.C. they
invaded Istria, regardlessof the fact that there was a very strong occupation
force in north-eastItaly, and even regardlessof the Romangarrisonin Siscia.
Local disturbanceswhich involved only the latter were probably a frequent
occurrence.The Scordisci,on the otherhand,wereon the decline.This led them
to form an alliancewith any neighbourwhich happenedto be strongerandthen
to play the part of mercenaries.They were, in fact, Burebista'smercenaries,and
in 16 B.C., as the allies of the Dentheletae,they plunderedin Macedonia.The
Romans obviously soon recognizedthat, as neighbours of the Pannonians,
Dardaniansand other tribes in westernThrace, the Scordisciwere a potential
dangerboth to Macedoniaand to northernItaly; but since their serviceswere
for sale, they were not difficult to win over to the Romanside.
In the years following Tiberius' campaignagainst the Scordisci there are
reports of renewedrevolts by the Pannonians,uIn 13 B.C. Agrippa himself
had to take the field againstthem, and after his deathTiberius was given com-
mand.I2 The presenceof Augustusin Aquileia in 12 B.c,I3 is a measureof the
importancehe attachedto theseoperations.Tiberius with his allies, the Scord-
isci, attackedruthlessly on this occasion.There are only extremelyscantyand
summaryreportsof the very tough fighting-magnumatroxquebellum14-which
ensued.From theseonly the following facts emerge,that one of the Pannonian
chieftains was called Bato-a very common Pannonianname-thatthe main
enemieswerethe Breuciandthe Amantinion the lower Save,andthat the battle-
groundwasinter Savumet Dravum.After the victoryTiberius hadthe adolescents
sold as slaves,an inhumanmethodof pacificationwhich was seldomappliedby
Augustus.Tiberius receivedthe triumphal insignia,I5
In the following years the Pannoniansagain resortedto arms. Up to 8 B.C.
Roman generals,in particular Tiberius, had to go repeatedlyto Pannoniato
attempt somehowto subduethe rebels who were waging guerrilla warfare.I6
Meanwhilein I I B.C. an imperial province under the nameIllyricum was con-
stitutedto includethe enormousareaof what was later to becomeDalmatiaand
Pannonia;the Danubewas declaredits northernboundary,17
In his personalaccountit is probably no accidentthat Augustus mentions
the Danube nor that he states that it has becomethe imperial frontier. He
34
consideredthe Elbe and later the Rhine and the Danubeas the only suitable
northernfrontiers of the empire. It was not that he had a preferencefor rivers
as clear dividing lines, or that as a southernerhe held rivers in specialregard,
or considerationsof that nature.Nor was it becausein his day thesebig rivers
would havebeenregardedas impossibleto cross-indescriptionsof the Celtic-
Germanic north there is abundantevidenceof the Rhine and Danubebeing
crossedin both directions:Augustusclearly recognizedthat not only the secur-
ity of the imperial frontiers but also the consolidationof Romanrule north of
the Alps could be achievedonly if Romeheld Europe'stwo greatrivers. More-
over he was fully aware that transport of goods and army supplies,and safe
communicationsin general,could only be achievedalong waterways.The trans-
Alpine provincescould not, however,be reachedeasilyfrom Italy, andcertainly
not by water. If Rome was interestedin trans-Alpineareasit was self-evident
that the rivers linking theseareasmust be in Romancontrol.
Within this contextit was only of secondaryimportancethat the Danubeand
Rhine should also become imperial frontiers. The campaignswhich soon
followed the constitution of the province of Illyricum are evidence that
Augustuswas intent on ensuringRome'sinfluenceon both banksof the Danube
and therebyprimarily on turning the river into an imperial traffic route. Shortly
after Illyricum was establishedas a province the governor Domitius Aheno-
barbusadvancedto the Elbe and also effectedchangesin the balanceof power
among the Germans by granting the Hermunduri new settlement areasI8
vacatedby the Marcomanni,whosemigration to the eastprobably occurredat
that time. The Marcomannihad driven out the Boii who lived north of the
Danube,and as a consequencethey remainedfor centuriesthe neighboursof
Noricum and Pannonia.I9 The Romanshad come to some sort of agreement
with the Marcomanni which, amongst other results, made trade relations
possible.20 In the following period the Hermunduri,as is well known, were on
good termswith Rome.21 Furtherto the eastthe Quadihad beengrantedsettle-
ment areas,probablyalso during the last decadeB.C. Their easternneighbours,
the Celtic Cotini, Osi, Taurisci and Anartii also cameinto contactwith Rome
at the time when Illyricum was being constituted.At this period they were
probably allies or subjectsof the Daciansafter their defeatby Burebistahalf a
centuryearlier.In 10 B.C. therewas a report of an incursionby the Daciansover
the frozen Danube;its repulseis actually mentionedby Augustusin his Monu-
mentumAncyranum.22 The Roman general on this occasionwas probably M.
Vinicius; a fragment of the inscription honouring his exploits, which was set
up in Tusculum,containsdetails of a campaignin which the Daciansand the
Bastarnae,their mercenaries,were defeated.Vinicius then advancedas far as the
35
Celtic raceson the northernedgeof the Hungarianplain and forced them into
an agreementwith Rome.23 According to Florus and Velleius he was engaged
as early as 13 B.C. in suppressingthe revolt of the Pannoniansand so was one of
Augustus' specialists in Danubian affairs.24 Another general, Lentulus, was
entrustedwith the task of bringing order to the southernedge of the Car-
pathianarea. His campaign,for which a precisedate cannotbe given, secured
the Danubefor Rome againstthe Daciansaround the Iron Gates.Moreover,
accordingto Florus, he was also successfulin keepingnot only the Daciansbut
alsothe Sarmatians'at a distancefrom the Danube'.LentulusestablishedRoman
guard-postson this side of the river.25
Although the precise years in which Lentulus' campaignstook place are
unknown (the dates given for them lie between 10 B.C. and A.D. II), their
context within the framework of Augustus'policy for securingthe Danubeis
clearfrom the brief descriptiongiven by Florus. In emphasizingthat the enemy
(the Daciansand Sarmatians)were 'kept at a distance'he addsthat this satisfied
Augustus. This is the only referenceto the establishmentof guard-posts(prae-
sidia) on the sectionof the Danubeoppositethe Dacians.There are reports of
severalDacian wars under Augustus,but neither their numbernor their dates
are certain.26 Whereasthe Germanand Celtic neighboursof the newly estab-
lished Danubefrontier were easily pacified either by military or by diplomatic
means,it was not so easyto establishpeacefulor friendly relationsbetweenthe
Daciankingdom and Rome. That was probablythe reasonfor placing garrisons
on the frontier, particularlythat sectionof it oppositethe Dacians.In fact in the
secondhalf of Augustus'reign there is frequent referenceto a Moesianarmy,
althoughat that time Moesiawas not yet a province.27 The commanderof this
army was subordinateto the governorof Macedonia.28
Within the Augustan system, which was followed by the Julio-Claudian
emperors,a very prudentbalancewas held betweendemonstrationsof strength
and diplomacy; wherever peacecould be achieved by diplomacy, that is by
alliancesand treaties,military measureswere deemedsuperfluousand avoided
as much as possible.Military occupationof the Danubefrontier at the time of
Augustuswas a last resort, and Lentulus was in fact the only generalto carry
out such a measure.'The watch on the Danube' even under Tiberius merely
meant that legions were stationed in a province which had the Danube
as a boundary, not that they were actually stationed on the bank of the
river.29
A further method of consolidation, often successfully adopted under
Augustus, was the resettlementof large ethnic groups. We have already seen
reasonto suspectthat the migration of the Hermunduriand the Marcomanni
36
was broughtabouteither at Rome'sinstigation or at leastwith her connivance.
In the first decadeA.D. Aelius Catus, commanderof the Moesian army and
possiblygovernorof Macedonia,had 50,000 Getaesettledsouthof the Danube
in what waslaterMoesia;30 the most significantmovementof this kind, however
was the transfer of the Sarmatiantribe of the Iazygesto the great Hungarian
plain. This settlement,so importantfor the subsequenthistory of the Carpathian
area, can be understoodonly within the context of Dacian-Romandisputes.
After Burebista'svictory over the Celts the Dacianswere in virtual possession
of the whole of the Hungarianplain. The Dacianwars underAugustuswere a
clear indication that the usual methods successfullyapplied elsewherewere
uselessagainstthe Dacians.The obvious thing to do was to interposea buffer
state in the plain, which would keep the dangerousenemy at bay. The first
general to come into contact with the Sarmatianswas Lentulus, who forced
them to respectthe Danubefrontier)! It is possiblethat Lentulus was in fact
the generalwho organizedthe transferof the Iazygesto the Hungarianplain. At
the time of Vinicius' campaign,at any rate, they were not yet settled in the
Carpathianregion,32but by the middle of the first centurythey are mentionedas
being old neighboursof the Quadi.33 According to Pliny they drove the Dacians
out of the plain into the mountainson its northernedge;34 their earliestsettle-
ment area was betweenthe Danubeand the Tisza35 and thus they were in a
position at least to keep the Daciansaway from Pannonia.
The great campaignto secure the Danube was launchedby Augustus in
A.D. 6, when he dispatchedTiberius with a very strong force againstMarobo-
duus, king of the Marcomanni.36 This campaignwas intended to pacify the
strongestGermanpower north of the Danube. Tiberius was already deep in
enemycountry when thePannonianson the initiative of Bato, chieftain of the
Pannoniantribe of the Daesidiates,revolted. Tiberius returnedimmediatelyto
Pannoniaafter he had concludeda peace-treatywith Maroboduus,which the
latter respecteduntil he was overthrown.37
There now followed three difficult years for Roman rule, so recently estab-
lished in Pannonia.38 As the causeof the revolt CassiusDio cites the masslevy
of the Pannoniantribes imposedby the governor, Valerius Messalla.In fact,
up to A.D. 6 Pannonianshad not been enlisted in the Roman army, not
even after the great war of 12-9 B.C.: at that time, it will be remembered,
Tiberius had hadthe young men sold as slaves.In the yearsbetween,boys who
had escapedslavery by reasonof their youth were growing up and, according
to Velleius, they gavethe Pannoniansconfidenceto throw off the unaccustomed
foreign yoke.
The courseof this war is better documentedthan that of Tiberius' war of
37
conquest,particularly as Velleius, an eye-witness,and CassiusDio both supply
equally detailed, though in many respectsdiffering, versions of it. Neither
source, however, gives any useful information concerningthe critical years
A.D. 7-8. What took placein Pannoniaafter A.D. 6 and beforethe surrenderon
3 August A.D. 8 can only be understoodif we visualize the rebels' guerrilla
tactics and Tiberius' clever and cautiousgeneralship.
All the Pannoniantribes both north and south of the Savejoined the rebel-
lious Daesidiates.In the north the Breuci, under their chieftain, who was also
called Bato, and in the south the Daesidiateswere the leading tribes. The
Pannonianswho at the outset were still a well-organizedand unified group
attacked Salona and Sirmium simultaneously:then Bato of the Daesidiates,
after being defeatedby Messalla, joined forces with the BreucanBato. Soon
CaecinaSeverus,commanderof the Moesian army, came to the rescueof the
Romangarrisonin beleaguered Sirmium anddrovethe rebelsinto the mountains
north of the city (Alma Mons-FruskaGora). From this point onwardsreports
becomeobscure.Leaving aside events in the area later to becomeDalmatia,
mention shouldfirst be madeof an attack on Macedonia,led perhapsby Das-
menus, which was probably repulsed by the governor Sextus Pompeius.39
CaecinaSeveruswith his army, now reinforcedby Thracianauxiliaries, had to
returnto Moesiabecauseof an invasionby the Daciansand Sarmatians.Tiberius
withdrew to the south-westcornerofPannoniaand establishedhis headquarters
at Siscia, on the Save,a basewhich had provedits usefulness.
In the following year (A.D. 7) there are reports of advancesunder Caecina
Severusand Plautius Silvanus, the latter having meanwhile been sent to the
aid of Tiberius; on the other hand, we hear reports that Augustus was dis-
satisfied with Tiberius' strategy of delay and as a result that, in addition to
Plautius Silvanus and others, Germanicuswas ordered to the seat of war.
Tiberius had, however,clearly recognizedthe situationand adaptedhis tactics
accordingly. Already, while troops under CaecinaSeverusand Silvanus were
being assembledfor deployment,it becameclear that the rebelswere not to be
lured into openencounter.A surpriseattackby them west of the AlmaMons in
the marshesof the Hiulca palus (also known as the Volcaepaludes)almostled to
the defeatof the Romantroops. The result was that Tiberius withdrew his men
to a mountainridge betweenthe Drave and the Save-knownin his circle as
Mons Claudius4o-applied a scorchedearth policy and awaited the collapse
of the rebellion which was showing increasinglack of cohesion.Meanwhile
negotiationswith some of the Pannonianleaderswere probably being started,
for when on 3 August in the following year the Breuci were forced by hunger
and diseaseto capitulateon the river Bathinus (Bosna),41their chieftain, Bato,
38
Conquestqf the Danuberegion and Augustanfrontier policy
was rewardedwith the rank of king. This same Bato also capturedPinnes,
one of the rebel leaders,and handedhim over to the Romans.
But Bato'skingdomdid not last long. The otherBato, chiefof the Daesidiates,
attackedhim, and the Breuci who had meanwhilerisen in revolt handedover
their king. The revolt was crushedby Plautius Silvanus.42 Probablyit was in
this sameyear (A.D. 8) that Illyricum was divided into twO.43 The northernpart,
probably not officially called Pannoniauntil much later,44 had already been
pacified; in the southernhalf, Dalmatia, it was not until the following year
that Bato, that doughty and consistentenemy,was defeated.
After the rebellion was crushed,such Pannonianyouths as were capableof
bearing arms were recruited as auxiliaries and withdrawn from the province.
Somealae and cohortesPannoniorumwere raisedat this time and in addition no
fewer than eight cohortsfrom amongthe Breuci.45 The cohortesLatobicorumand
Varcianorum likewise probably now came into being. In the Julio-Claudian
period all Pannonianauxiliaries attestedeither on inscriptions or in military
diplomas come from southern Pannonia.46 Recruitment of auxiliaries from
amongthe northernPannoniantribes did not begin until later, roughly about
the middle of the first century.47It would seemthat only the tribes southof the
Drave took part in the rebellion: Suetoniussuggeststhat north of the Save
only the Breuci were involved.48 The next chapterwill show that after A.D. 8
Tiberius used very harsh measuresof pacification, which brought about the
disintegrationof the formerly powerful Breuci. Thus it is very probablethat
beforethat datethe Breuci, as the only rebelsin Pannonia, werein control of a
fairly big areain the Savevalley. In any case,not all the tribes evenin southern
Pannoniarebelled; the Scordisciseemto have remainedsteadfastin their sup-
port of Rome.49
In the last years of Augustus'reign there was peaceon the middle Danube.
Lentulus' war againstthe Dacians and Sarmatiansmay have occurredat the
latest around A.D. 10. There is much to be said in favour of this date, since
according to Florus, our main source, Lentulus settled both the Dacian and
Sarmatianproblems for quite a long time to come. Yet in A.D. 6, these two
tribes had attackedMoesia.50 Thus the settlementof the Iazygesin the Hun-
garian plain might well have taken place around A.D. 10. The date given by
most scholars,however,is 17-20, when Drususwas active in Illyricum. 51
The yearsof peacewhich followed A.D. 10 or I I broughtno illusion at Rome
that the military and diplomatic measuresundertakenacrossthe Danubehad
consolidatedthe situation in final form. Apparently Augustus wished first to
embarkon further stepssouthof the river with the aim of consolidatinghis rule.
In 14 he sent Tiberius to Illyricum adftrmandapace quae bello subegerat.52 Details
39
of this commissionare lacking as the death of the agedemperorshortly after-
wards forced his return. One of his tasks may be inferred from what followed.
On the news of the death of Augustus the legions in Pannoniarose up in
revolt.53 They hadfairly substantialreasons:accordingto Tacitus,the soldiers-
in the main old and about to be discharged-wereopposedto grants of 'un-
cultivated mountainousland and swamps' for veteran allotment. Now it is
known that the deductio of colonia Julia Emona(Ljubljana) was actually started
in that year. By the end of I4 the new town's wall was finished.54 The soil
round Emona was, however, far from suitable for agriculture; the Laibach
marshesreachedright up to the town-wall, and from the further edge of the
marshlandthere rose the spurs of the Julian Alps. At the sametime or even
earlier, legio XV Apollinaris was transferredto Carnuntum.55 From this it is
clear that Augustus wished among other things to start the colonization of
Pannoniaand simultaneouslyto transfera legion close to the Marcomanni.
Both tasks were carried out by Tiberius. His other work on the middle
Danube was no more than a logical, though less energetic, continuation of
Augustus' programme. When in 17 he sent his son, Drusus, with special
plenary powers to Illyricum,56 the latter's tasks were probably not basically
different from those which his father should have carried out in 14. The sub-
sequentperiod throws light on the functioning of Augustus'Danubepolicy.
In 17, when a conflict was about to break out betweenMaroboduusand
Arminius, the former, relying on a peace-treatyconcluded with Tiberius,
appealedto Rome for help. The plea remainedunanswered.Instead Tiberius
had Maroboduusoverthrownby Drusus.The Germanking fled to the empire,
where he was grantedasylum and internedin Ravenna.Shortly afterwardsthe
new king, Catualda,who had beeninstalled by Drusus, was also overthrown,
and he, too, was given asylumin the empire.A new political structurewas then
formed north of Noricum and Pannoniaunder Vannius, king of the Quadi,
whose rule lasted for thirty years. Drusus, who knew how to exploit these
upheavalsin Rome'sinterest,celebratedan ovation in Rome in 20.57
The policy initiated by Augustusalso stoodthe test when Vannius was over-
thrown in 50.58 Romewas uninterestedin the internalpolitics of Vannius' king-
dom so long as he remaineda reliable vassaland his position was sufficiently
strong.Meanwhilehe had considerablyextendedhis empire eastwardsand had
madesomeneighbouringtribes,including possiblythe Cotini and Osi, liable to
tribute.59 He also made some kind of alliance with the Iazyges,for the royal
cavalry was recruitedfrom them. This extensionof power by a Romanophile
king was welcomed by Rome; but when Vannius' control began to wane
becauseof a palacerevolution, the EmperorClaudius was concernednot with
40
supportinghim but with winning over the leadersof the revolt to the Roman
cause.PalpelliusHister, governorof Pannonia,was orderednot to intervenein
the civil war ragingin Vannius'kingdom, but to recognizethe latter'srebellious
nephews,Sido and Italicus. Along with his followers Vannius was settled in
Pannoniaand his nephewsenteredinto their inheritanceas vassalsof Rome.60
The combinationof methodspursuedby Augustusin his Danubepolicy is
seenparticularly clearly in the activities of a Moesian governor.In the second
half of Nero's reign Plautius SilvanusAelianus had more than 100,000barbar-
ians settled in the province, intervened militarily in a Sarmatianmovement
which was under way, forced princes of the Dacians,Bastarnaeand Roxolani
to make submissionbefore Roman standards,took hostagesand confirmed
certainkings in power. For his serviceshe later receivedthe triumphal insignia
from V espasian. 61
This successfulDanubepolicy lasted until the Year of the Four Emperors.
So long as there was no doubt that Rome was in a position at any momentto
intervene with force or diplomacy in the affairs of the peoples beyond the
Danube,it was not difficult to maintaina balance.In the confusionof the years
68-9 the best troops were, however, withdrawn from the Danubeprovinces
and in various placesa power vacuumensuedbetweenthe rival Romanparties.
The governorof Pannonia,TampiusFlavianus,the legionarycommanderswho
met in Poetovio, and the procuratorof the province, Cornelius Fuscus,were
still able to apply Augustus' methods until, along with the Pannonianarmy,
they becameinvolved in the civil war. But in Moesia, which had beendenuded
of troops and which, of course, confrontedthe Dacians (among whom little
sign of submissivenesswas to be found), the old methods completely co1-
1apsed.62
Once again the exampleof Pannoniarevealshow thesemethodsfunctioned
in time of crisis. After it had beendecidedin Poetovioto march with the Pan-
nonian legions to Italy againstVitellius, various measureswere taken to deal
with the situation beyondthe frontier. Sido and Ita1icus, princes of the Quadi,
and the princes of the Iazyges not mentionedby name, were summonedto a
conferenceto discuss the furnishing of military aid in accordancewith the
existing client treaty. When the 'discussion'ended,the princes of the Iazyges
were detainedas hostagesand the troops they had brought with them were
sent home; Sido and Ita1icus, togetherwith their troops, were taken to Italy.
The difference in the treatment of the two client-statesis evidence both of
familiarity with the situation and of mature judgment of it. Sido and Italicus
were loyal to Rome, but were hatedby their own people.If the troops which it
was their boundenduty to provide had beentaken to Rome without them, the
4I
two princes could easily have been overthrown. Equally, to have sent them
homewith their troops would haverisked subjectingthe Germansto a tempta-
tion which perhapsthey might not have beenable to resist. In the caseof the
Iazygesit was sufficient to hold their leadershostage;it was impossibleto say
how their troops would behave in the approachingconfusion. That these
measureswerebasedon sound judgmentis indicatedby the peacefulconditions
which prevailedin Pannoniaduring the period of crisis in contrastwith thosein
Moesia,which fell to the attacksof the barbarians.For his diplomatic successes
TampiusFlavianusreceivedthe triumphal insignia from Vespasian.63
The situationin Moesiabecamecatastrophic.The Roxolani, a tribe of horse-
men who had recentlyappearedon the Danubefrontier and with whom as yet
the Romans had not been able to establisha client-relationship,invaded the
provincein the winter of 67-8. Despitethe destructionof two cohorts,M. Apo-
nius Saturninus,the governor,managedto repulsethem.64 Becauseof his ambi-
valent attitude during the civil war this governorhad to leave the province in
69.65 His legions went to Italy and the Dacians were able to invade Moesia
acrossthe unprotectedDanube.It was by good fortune that Licinius Mucianus
happenedto be on his way to Italy from the eastwith a legion, and successfully
repulsedthe invaders.66Shortly afterwardsthe provincewas againinvaded,this
time by the Sarmatians,who defeatedthe new governorFonteiusAgrippa.67 It
wasnot until 70 that his successor,Rubrius Gallus,wasable to establishpeace.68
It was this combinationof successin Pannoniaand failure in Moesia on the
part of the Augustan systemthat persuadedthe new Flavian governmentto
replace it on the Danube with a new frontier policy. Rubrius Gallus had
alreadybegunthe constructionof fortressesalong the river, and soonthereare
reports of similar stepsin Pannonia.Under the Flavians there gradually deve-
loped a fundamentally new strategic concept. It was not considerationsof
foreign policy alone which brought about this change.The occupationarmies
in the provinces, whose original task had been to keep the peacewithin the
province ratherthan to fight enemieswithout, graduallylost their former role.
It was an importantlessonof the Year of the Four Emperorsthat it was better
not to stationlegions in pacified areasclose to Italy, but rather on the frontier.
It was this recognitionwhich brought the first period of occupationto an end.
The earliestdistribution of the considerablenumberof legionssentby Augustus

to Illyricum is as good as unknown. Deductionscan only be madefor a later
period, but it is neverthelessnot improbable that the centres of occupation
remainedessentiallythe samefrom the period of conquestup to roughly the
time of Claudius.In the northernpart of Illyricum (which becamePannoniain
42
A.D. 8) therewere threelegionsunderAugustus,VIII Augusta,IX Hispanaand
XV Apollinaris. Epigraphicevidenceshows that VIII Augustawas in Poetovio
from at leastthe time of Tiberius.69 By 141 15 at the latest legio XV Apollinaris
was stationedat Carnuntum.Up to the presentit has not beenpossibleto estab-
lish whereIX Hispanawas based,nor indeedwhethereachlegion had its own
fortress. In any case neither Poetovio nor Carnuntumcan be regardedas a
double fortress, for the legionariesleft many inscriptions in both places,yet
neitherplacehas yielded any tracesof IX Hispana.It is usually assumedthat it
lay at Siscia, the most importantstrong-pointin the Savevalley, and one which
Romantroops had occupiedsince 35 B.C.
At the time of Augustus' death thesethree legions were concentratedin a
summercampand had to carry out various kinds of constructionwork, mainly
road building. This camp was probably in the upper Save valley, where one
detachmentbuilt roads and bridges around Nauportus. Apart from this,
Tacitus' description of the legionaries' uprising says nothing to further our
knowledge.70 It is frequently assumedthat under Augustus legio XV Apol-
linaris was stationedin Emona but was transferredto Carnuntumbefore the
colony wasfounded.Thereis somethingto be saidfor doubtsrecentlyexpressed
whethertherewasa legionaryfortressat Emona.71 It is not impossiblethat legio
XV Apollinaris had movedto Carnuntumat an earlier date,roughly at the time
when the Marcomannioccupiedthe northern part of what had formerly been
Boian country. Apart from Carnuntumit must be assumedthat underAugustus
therewas a strongoccupationforce in Siscia,andlikewise in Sirmium, wherein
A.D. 6 its Romangarrisonwas attackedby Pannonianinsurgents. 72
The Moesian army-command,which was subordinateto the governor of
Macedonia,was not organizeduntil after the conquestof Pannonia.It also
consistedof severallegions, and in particularprobablyincludedV Macedonica
and IIII Scythica.The former, like XV Apollinaris, was stationed,at leastfrom
the time of Tiberius, on the Danubeat Oescus(Gigen), whereit had a number
of auxiliaries alongsideit.73 The epithet 'Macedonica'suggestsperhaps that
underAugustusthis legion had beenstationedfor a time in Macedoniaor north
of it; but the possibility cannotbe dismissedthat the legionaryfortressat Oescus
was establishedas early as the periodwhenLentulus,CaecinaSeverusandAelius
Catuswere active on the Danube.The building of garrison-postson 'this side
of the Danube' by Lentulus74 may easily therefore refer to the transfer of V
Macedonicato the river. The two legions which were moved forward to the
Danubeeach kept a barbarianstate in check: XV Apollinaris at Carnuntum
watchedthe kingdom of Maroboduusand V Macedonicaat Oescuswatched
the Dacians,who until Lentulus' campaignshowedlittle sign of submission.
43
Internalconditionsimprovedto suchan extentin Pannoniaduring the decade
and a half after the crushing of the rebellion that legio IX Hispana could be
temporarily transferred to Africa against Tacfarinas. At the time, Tiberius
reportedto the Senatethat the Danubewas being guardedby two legions in
Pannoniaand by two in Moesia.75 It was about this time that the latter was
constituteda province. While no datecan be given for this event,Appian states
explicitly that Tiberius annexedit to the empire.76 The first governorofMoesia,
PoppaeusSabinus,was at the sametime governor of Achaeaand Macedonia,
and when he died after twenty-four years' service in the Balkans the two un-
armed Greek provinces and Moesia, which was occupied in strength, were
againentrustedto a single governor,Memmius Regulus.ConstitutingMoesiaa
province was thus probably only a formality and was not precededby a war of
conquest.There was no changein the commandof the forces stationedin the
province: their commanderwas responsibleas legatuspro praetore to Poppaeus
Sabinus.77
It would seemthat Tiberius made no fundamentalalterationin the deploy-
ment of troops which had beenarrangedin broad outline by Augustus: Pan-
nonia and Moesia eachhad a legion on the Danube,there were two in the in-
terior of Pannoniaand onein the interior of Moesia.Where legio nn Scythica
was stationedis still an open question.Most probablyit was somewherein the
south, close to Macedonia; Scupi springs first to mind as there is epigraphic
evidencethere,78or Naissuswhere auxiliaries are attestedin the first century.79
Both placeswere important strategicpoints in the link betweenMacedoniaand
the Danube.
This route was far from easyto negotiate.The Vardar ravine on the Mace-
donian borderinitially presentedan obstacle,while the narrow gorges(Serbian
Klisura) in the Morava valley north and south of Naissusand in the Timok
valley madethe deploymentof troops difficult (PI. 7). It was in fact as difficult
to reach the Danubefrom Macedoniaas it was from Italy. There was a much
easierroad from Macedoniathrough the Struma (Strymon) valley via Serdica
andthenthroughthe Isker valley. Crassushadalreadyusedit in 29 B.C. This road
terminatedat Oescus.
In view of the difficulties of communicationTiberius vigorously set about
building roads. In the summerof A.D. 14 Pannonianlegions had alreadybeen
assembledfor this purposeand for the constructionof bridges in the upper
Save valley.80 It obviously took several years to complete the roads from
Aquileia to Carnuntumon the Danubeand as far as the mouth of the Save.
During Drusus' administrationin Illyricum further roads were built; several
connectedSalonato the Save;that this was a natural direction for them to take
44
is attestedby the distribution of Republicancoins (Fig. 7).81 The building of the
road betweenSalonaand Sirmium in 19-20 provideda secondlink betweenthe
central position of Sirmium and the Adriatic.82
Theseroads,radiatingfrom the Mediterraneanto the Danube,providedonly
overlandlinks which, comparedwith other possibilities,were alwaysof second-
ary importance. Communicationby water, as has been seen, could only be
achievedby linking the provinces along the Danube. This becamenecessary
at an early stage,sinceit was easiestto supply the legions stationedon the river
at Carnuntumand Oescusby water. Either Augustusor Tiberius developeda
Danubefleet. The first mention of it is in 50; 83 however,it must certainly have
beenin existenceby the end of Tiberius' reign as his most outstandingachieve-
ment in the Danube provinces,the building of the cliff road in the Djerdap
(Iron Gates)(Fig. 8), is impossibleto imaginewithout a fleet. Just as in the last
years of the reign of AugustusPannonianlegions were usedto build roads in
the upper Savevalley, so the two Moesianlegions were concentratedin 33-4 in
the ravine above the Iron Gates for the purposeof carrying out large-scale
constructionwork.84 Once the radial roads to the Danubewere finished, their
termini had to be connected.This done, work on the Danube frontier was
complete.
In order to appreciateproperly this colossalundertakingit is necessaryfirst
to give a brief descriptionof this sectionof the Danube.85 Whereit cuts through
the Carpathianand Balkan massif it flows for about 130 km through a very
narrow gorge; in the Carpathianbasin the river in someplacesis 600 m wide
and occasionallyeven wider, but it then forces itself through a narrow valley
which in the Kasangorgeis reducedin placesto a width of only 150 m (PI. 8b).
In fact it is only near Donji Milanovac, where the Poreckaflows into it, that
there is any openingto the south; otherwisethe southernbank is mostly pre-
cipitous and unapproachable.There is no natural road along this side. In the
Kasangorge,which is roughly 40 km long, the rock-faceis so closeto the river
that ships cannoteventie up there (PI. 9a). The rocky river-bedis an additional
obstacle.An outstandingfeature of the section above Donji Milanovac is the
rapids, though the biggestare near the Iron Gatesbelow OrSovaat the end of
the Kasangorge. The Romansprobably called the rapids in the upper Djerdap
(Gornja Klisura) scrojulae,86'little pigs', a term for rock-boundwaterswhich is
also found in other languages.In ancientliterature the word 'cataracts'is also
used of the Djerdap section of the Danube.87 Finally, whirlpools and the fast
current make navigationvery difficult in this stretchof the river.
Construction under Tiberius was directed towards making this section
generallynavigableand to linking the upperand lower Danubevalleys. It is no
45
Figure 8 The tow-path in the Djerdap (reconstructionby E. Swoboda)

Top, as originally constructed.
Bottom, as reconstructedunder Trajan.
46
longer possibleto say whether and to what extent the rocks in the river-bed
were removed.It is not true that all this work had to be repeatedat the time of
the extensiveworks to straightenthe river-bedin the nineteenthcentury,for the
simple reasonthat the work carried out then was to meet the needof steamers.
In many respectsthe nineteenth-centuryworks do, however, tally with the
Roman,as E. Swobodapoints out in his authoritativebook. Particularlywhen
travelling up-stream,ships had to be towed either by men or animals (on the
new Iron Gatescanal enginesare used). In the Djerdap towagewas, however,
impossible.First of all a tow-pathhad to be built, a task which calledfor all the
technical skill of antiquity. It meant either that a road had to be cut into the
precipitouswall of rock, often of the hardestkind imaginable,or at leasta path
resting on supportshad to be constructed.According to two inscriptions in
the upper Djerdap, the task was completedby legiones nn Scythica and V
Macedonicain 33/4.88 Otherinscriptionsindicatethat work wascontinuedunder
Claudius (when the same legions were involved), and under Domitian and
Trajan.89 The Domitianic inscriptions refer to the road as iter Scorfularum or
Scrofularumand indicate that partly through use and partly owing to flooding
the road had to be rebuilt. Trajan's Tabula in the Kasangorge (Pl. 9b) encour-
agedscholarsfor a long time to regardthe tow-paththereas a new construction
concernedwith his Dacianwars. The road was, however,only useful if it made
navigationpossiblethroughthe whole of the Djerdap,from Golubac(Cuppae)
as far as Orsova (Dierna). A new reading of the Tabula by E. Swobodaulti-
mately proved that Trajan only rebuilt the road (viam refecit). This rebuilding
must not, however,be underestimated,since,accordingto the text, it was then
that the road was first cut in many placesinto the cliff walls; as the supportsand
joists of the earlier road were now superfluous,they were removed.The holes
hewn at regular intervals into the rock below the road and into which its sup-
ports were fixed can still be clearly seenat many placesin the Djerdap. But this
later road, which in places was cut into the rock to a depth of 2-3 m, also
requireda woodenconstruction.It is obvious that sucha road neededconstant
maintenanceand protection.Otherwise,on this stretchof the Danubetherewas
hardly any dangerfrom Daciansand Sarmatians,as the towering cliffs, and the
very steep slopes on the river's southernbank, doomedany attack to failure
from the start. The chief danger came from the elements,since frequently at
the end of the winter ice piled up so high in the narrow gorgesthat the wooden
supportswere carriedaway with the melting waters.Therewas also the further
risk that maraudingbarbarianswould set fire to the supportsand thus interrupt
suppliesfor the fleet and troops. It was obviously to counter this dangerthat
Tiberius had alreadystationedauxiliaries on this sectionof the frontier. But on
47
Conquestoj the Danuberegion and Augustanfrontier policy
a section 130 km long there were only a few places suitable for normal-sized
forts. It can be assumedthat evenin later times there were auxiliary forts only
at Golubac at the entranceto the Djerdap, at Cezava(Novae) and at Donji
Milanovac (Taliata).9o The remaining guard-postswere fortlets and watch-
towers.91
The dispositionof troopson the middle Danubewas modified underClaudius
to the extent that legio IX Hispana was moved from its Pannonianfortress
(Siscia ?) never to return, and one legion each in Pannoniaand Moesia was
replaced.Legio XIII Geminatook the place of VIII Augustain Poetovioand
IIII Scythica was replacedby VII Claudia, which was thereafterpermanently
basedin Moesia. In Pannoniafortressesremainedunchanged(Carnuntumand
Poetovio); whetherVII Claudia moved into a new basein Moesia or into the
(unidentified)fortress of its predecessoris so far impossibleto say. The theory
is frequently entertainedthat it moved to its later baseat Viminacium under
Claudius, but in the presentstate of our knowledgeit is almost impossibleto
show that a legionary fortress existed at Viminacium, Naissus,Scupi or any-
where else in Moesia in the first century.
Under Nero two legions were stationedin Pannoniaand one in what was
later Upper Moesia. In all the Romanperiod this was the time when the occu-
pation army of Pannoniaand Moesia was of least importance.It seemedas if
hereRomehadno far-reachingprovincial plan apartfrom guardingandsecuring
the Danubefrontier, that eventhe requirementsof frontier policy wererestricted
to a minimum. Nor did the transferof troops in the final years of Nero's reign
bring about any appreciablechanges.Legio XV Apollinaris, dispatchedto take
part in his war in the east,was replacedby X Gemina,then by VII Geminaand
also by XXII Primigenia.
The questionnow arisesof the distribution of the not inconsiderableauxiliary
forces92 in theseprovincesin the Julio-Claudianperiod. It is only when this is
appreciatedthat the strategy of Roman policy on the Danube and sensitive
elementsin the internal situation becomeclear. For a long time it was difficult
for scholarsto realize that althoughthe Danubehad beendeclaredan imperial
frontier under Augustus it is not until later that its military occupation is
attested.Mommsen and subsequentlymany others interpretedprotulique fines
imperii usquead ripam jluminis Danuvii as referring to the short stretchof the river
from the mouthof the Draveto that of the Save93 andassumedthat the conquest
of the areato the north of the former river was a gradualprocess.Even after it
had beenestablishedbeyonddoubt that a legionaryfortress had beenset up at
Carnuntumat the latest under Tiberius it was still believed that the conquest
of the northern and easternparts of Pannoniawas postponedto the period
48
C01tquestoj the Danuberegion and Augustanjrontier policy
betweenthe reigns of Augustus and Vespasian.The Rhine frontier, with its
strong concentrationof troops and camps,can hardly be cited as comparable,
for the political situation on that river was completely different. From an
administrativepoint of view it is, however,probablethat someauxiliary troops
were sent forward to the Danube:there are also direct, though scanty,indica-
tions that there were Julio-Claudianauxiliary forts there.
As for the administrationof the newly conqueredprovince,it will be seenin
the next chapter that the indigenous communities (civitates) came under the
military control of a high-rankingofficer, frequentlya memberof the Equestrian
Order,from a neighbouringRomanunit. We would probablynot be far wrong
in assumingthat in the Augustan-Tiberian period there was at least one auxi-
liary unit for every two civitates.The eight civitateswhoseterritories extended
to the Danubethus requiredat leastfour auxiliary regiments;it is probablethat
legio XV Apollinaris controlledsomeof the north Pannoniancivitates,andthat
thereforethe numberof auxiliary troopstransferredto the Danubefor internal
political reasonswas not particularly large. The samewas true in Moesia, only
there it was probably not until the time of Tiberius that there was consistency
in the military administration.
Neitherin Pannonianor in Moesiahavearchaeologicalexcavationsproduced
evidenceof pre-Flavianauxiliary forts. Perhapsthe only exceptionis at Adony
(Vetus Salina) south of Budapest,where a short-lived fort may possibly have
existed in the middle of the first century.94There have been fairly extensive
excavationsof the forts of the later limes in Hungary, particularly between
Arrabonaand Intercisa,but nowhere,apartfrom Adony, is there any evidence
that can possibly be dated to the pre-Flavianperiod. Either pre-Flavianforts
were not sited on the Danubeor later forts were not built on pre-Flaviansites.
At Vetus Salinafort II, which is possiblypre-Flavian,was discoveredsouthof
its successor.More recently, excavationsin Szazhalombatta(Matrica) and at
Intercisahave revealedtracesof fort ditches outsidethe areaof the later sites,
but as yet they have not beendated.95
Smallfinds which canbe dated,suchas pre-Flaviansamianor coins,yield even
less conclusiveevidenceof auxiliary forts on the Danube.More reliable evi-
dencefor such sites at somepoints along the river comesfrom tombstonesof
pre-Flavianauxiliary soldiers and veterans.Such tombstonesare unknown on
the Moesianfrontier but are numerousin Pannonia.The fact that theseauxiliar-
ies were almost without exceptiontroopersbelongingto various alae is prob-
ably the result of a tactical principle of the early occupationperiod. On the
evidenceof tombstonesthe following placesalong Pannonia'sDanubefrontier
can be regardedas sites of pre-Flavianauxiliary forts: Arrabona,96Brigetio,97
49
Aquincum,98Lussonium,99Teutoburgium100 and MalataIOI_a chain of pretty
evenly distributedsites. The questionnow arisesas to whetherthey all existed
under Augustus and Tiberius. In view of the complete lack of supporting
archaeologicalevidence,the view here suggestedis that most of the fort-sites
listed and possibly othersat presentunknownwere not establisheduntil about
the middle of the first century. Proof of this is, of course,impossible.At Arra-
bona,whereala I AugustaItyraeorumand ala I Pannoniorumleft behind many
pre-Flaviantombstones,a recent study of early Roman finds 102 suggeststhat
the site was not establisheduntil after the middle of the century; at Brigetio,
where Nero's praetorianprefectC. Nymphidius Sabinusaspraefectusalae set up
an altar,103 the evidence of samian likewise begins after the middle of the
century;104 at Aquincum in the fort of ala II Asturum and ala I Hispanorum
early Romanarchaeologicalmaterial goes back to the sameperiod.lOS
Whether the establishmentof theseforts was an innovation on the part of
Claudius it is impossibleto say. Be that as it may, he was the emperorwho
foundedthe first colonia north of the Drave; moreoverthe first definite indica-
tions that auxiliarieswere being recruitedin northernPannoniaalso date to the
Claudianperiod.106
The main body of the occupationforce in pre-Flaviantimes was concentrated
either with the legions or in the interior of the province. Here the problem is
evenmore difficult than it is with the occupationforces on the frontier. At least
we know that a significant number of the auxilia were accommodatedin or
aroundthe legionaryfortresses;they are, however,epigraphicallyattestedonly
at Carnuntum,andeventhereonly in small numbers.107In the earlyperiodit was
exceptionalfor auxiliary soldiers to set up tombstones:they probably did so
only where there was an active stonemason'syard or where one within easy
reach could deliver the finished product. Stonemasonswho could chisel a
tombstone,and especiallythose who were specialistsin that field, were to be
found in pre-Flaviantimes only at a few placesin western Pannonia,mainly
near legionary fortressesand at Emona, Savaria, possibly at Scarbantia,and
perhapsalso at Aquincum in easternPannonia.Except at Carnuntumtomb-
stonesof auxiliary soldiers have beenfound only at Aquincum, on somefort-
sites on the Danubeand in the vicinity of Scarbantia.Tombstonesof soldiers
belongingto ala I Hispanorumand ala I Pannoniorum108 have beenfound at
Mattersdorf,north of Scarbantia,and at Gya16kato the southof it respectively.
It can be deducedfrom thesefinds that cavalryunits were stationedon the road
from Savariato Carnuntumin the middle of the first century.This stationingof
cavalry units on roadsleading to the imperial frontier is regardedas the most
important characteristicof the Julio-Claudiansystemof occupation.Whether
50
the recently discoveredearthworkfort at Gorsium (Tic) was also part of it has
not yet beenestablished.It is possiblethat it was garrisonedby the ala Scubu-
lorum.109 Apart from that at Gorsium, mention should also be made of a fort
discoveredby excavationat NadleSki Hrib, south-eastof Emona, which has,
however,not yet beenfirmly dated.Its site is closeto that of a hill-fort (Ulaka)
belonging to the native inhabitants and must have served to keep the latter
under observation.
In Moesiapre-Flaviantombstonesof auxiliary soldiershave beenfound only
at Naissusand possibly at Ratiaria. That at Naissus commemoratesa Cretan
soldier who had probably been dischargedunder either Claudius or Nero,I10
The stoneat Ratiaria might alternatively be of Flavian date.III In any casethe
former allows us to infer that auxiliary units were stationedon the roadthrough
the Moravavalley in pre-Flaviantimes. North of Naissus,nearRavna(Timacum
minus) in the Timok valley there is epigraphic evidenceof an early Flavian
auxiliary fort; II2 it is highly probablethat it had a predecessoron the samesite.
Place-namesin the Morava valley, such as HorreumMargi and perhapsPraesi-
dium DasmeniandPraesidiumPompeii,suggestthat therewas a roadtherewith
guard-postsand supply depots.If legio nn Scythicaand also perhapstempo-
rarily vn Claudia were stationedin southernMoesia (at Scupi?or Naissus?)it
is reasonableto supposethat auxiliary units held a road through the valleys of
the Morava and the Timok.
The auxilia attachedto the army of occupationin Pannoniawerealreadybeing
augmentedin the Julio-Claudianperiod by local recruits. Tacitus makes special
referenceto this with regard to the year 5o.II3 Some tombstonesreferred to
above are those of south Pannoniancavalrymen. Tombstonesand military
diplomasclearly indicatethat the civitates on the Saveand Drave providedthe
army with a very largenumberof auxiliaries.The Breuci, Colapiani,Cornacates,
Sisciani, Varciani, Iasi and Latobici, all of them south Pannoniantribes,II4
are mentionedby name.Only the first four of theseare attestedas contributing
to the Pannonianarmy; II 5 the young men from the others were used outside
the province.This patternmay be dueto fortuitous transmissionof information;
neverthelessit is striking that the Colapiani lived in the Kulpa valley in the
vicinity of Siscia and had thereforebeen under Roman control since 35 B.C.;
Siscia had also beenTiberius' headquartersnot only in II-9 B.C. but again in
A.D. 6-9, so it may be assumedthat the Colapiani,who were amongthe earliest
nativesto be pacified, were also amongthe most reliable. The samewas true of
the Varciani, westernneighboursof the former. The gradualextensionof local
recruitmentof the auxilia is easyto trace. By the middle of the first centurynot
only the lasi but now north Pannonians,too, were being recruited into the
51
Conquestof the Danuberegion and Augustanfrontier poliC)I
Pannonianarmy.II 6 It is possiblethat more Pannonianswere being draftedinto
the navy, since in the Year of the Four Emperors thefleets of Ravennaand
Misenum were both largely mannedby Pannoniansand Dalmatians.II7Tribes
which were not yet thoroughly pacified were probably consideredsuitable
materialfor the navy. That the gradualextensionof recruitmentareasis a pal-
pable symptom of the consolidationof Roman rule is proved by the fact that
until late in the secondcenturyUpperMoesianswere not only not calledup for
service within the province but were not used outside it. It should be borne
in mind that it was the governor'sorder for recruitmentwhich was the imme-
diate causeof the Pannonian-Dalmatian revolt in A.D. 6.n8 In Moesia, which
was never acquiredby conquest,drastic interventionwas avoided; in contrast
to the cohortesand alae Breucorum,Latobicorum, Varcianorum and Pannoniorum
therewas only a single auxiliary unit bearingthe nameof a Moesiantribe: this
was the ala VespasianaDardanorum,but even so it was a creationof the Flavian
period.
52
Chapter 3
Native population and settlement
The principalpeoplesandtribes of pre-Romantimesareall mentionedalsoin the

Roman period as inhabitants of Pannoniaand Moesia. Our most important
sourceis Pliny the Elder, who in Book 3 of his Natural History describesthe
provinces and lists their inhabitants. These lists, arrangedalphabetically,are
basedon a descriptionof the empire, begun by Agrippa and continuedunder
Augustus;there is everyreasonto believe that suchlists were the official regis-
ters of the civitates peregrinae. A later listing of indigenous tribes, arranged
geographicallyand probably also based on official sources,is the work of
Ptolemy; however,the probability is that it is neithercompletenor administra-
tively accurate.As far as Pannoniais concerned,thereis substantialagreement
betweenPliny's list andthat of Ptolemy;for Moesia,however,thereis a marked
discrepancywhich is not without a certainsignificance.
Apart from the peoples who were of importance before the province of
Illyricum was constituted (the Boii, Taurisci, Breuci, Andizetes, Amantini,
Scordisciand Latobici) (Fig. 9), Pliny's list! containsa whole seriesof peoples
who are mentionedfirst in imperial times, someof them only by him. It is con-
ceivablethat such peopleswere not included in descriptionsof conditionsand
eventsin pre-Romantimes because,being oflittle importance,they were under
the control of their more powerful neighbours.This may well have been true
of some,but is not generallyapplicable.A considerablenumber of peoplesin
Pliny's list haveLatin names,which suggeststhat they were in fact new groups
organized by the Romans. Some of these 'tribes' were called after places:
Cornacatesafter Cornacum(Sotin, on the Danube, east of Mursa), Oseriates
(place-namenot attested), Varciani after Varceia (a place which, although
53
Nativepopulation and settlement
^M
it
o , 100 ,.
1 «to
hr
^M ame
it u
hr
ame
u
^M
ith
eu
ra
eu
ra
ith
^M
^Mithr
a me u
^M
ith
ra
^M
u
^M
r a me
eu
h
Mit
it
it
h r^ M
hr
^M
amei t
ame
it
uhr
hr
u
u u
ame
ame
e
ae u
t h rae r a
thr i
u
u
i ^M u a me u h ^M eu
^M at me it i t h r a me u
r
a
h r h M u
aeeu thr
it i
^M ^M ^ r
h r a i
u it th ^M
r a me
eu
a e
M ui t h a ^ MM i
r^ m hr ^
eu
ith eu u u
^M
it
hra a ea e
a
e u
i t ^M
r â M it e
hr
r
ith h rh ar a
^
itith ^^MM i t
^M it
it
^M eu
h
Mi h r
^M
^ M^ M it thh a
rraa e u
eeu
u
Figure 9 The native peoplesof Pannonia
attested,2cannotbe located);otherswere called after rivers: Colapiani(Colapis,

Kulpa), Arabiates(Arabo, Riba, GermanRaab),To thesemay perhapsbe added
the Hercuniates,calledafter the Herrynia silva (this probablypreservesthe legend
that they camefrom the north), Apart from the Hercuniatesand the Arabiates
the peopleswith Latin nameslived in the Savevalley (Fig, II), Normal tribal
namesnot attestedin pre-Romantimes are to be found in the Dravevalley (the
Serretes,Serapilli, Iasi) and in northernPannonia(the Azali and Eravisci),3
It is well known that the tribeswho put up the strongest resistanceto Tiberius,
and who also led the Pannonian-Dalmatianrevolt, were located in the Save
valley; our sourcesmentionthe Breuci and the Amantini, But betweenthem in
Romantimes the Cornacatesappeared,a group which had without doubt been
organizedby the Romans, In the Save valley, west of the Breuci, were the
54
Nativepopulationand settlement
Oseriates,Colapiani and Varciani, all of them with Latin names.It would seem
that after the subjectionof the Pannonians,the extensivearea formerly con-
trolled by the Breuci and Amantini was dismembered,and some of the new
administrativeunits createdby the Romanswere either given the namesof the
old tribes, or, in most cases,were called after placesor rivers. This drasticinter-
ferencein the balanceof poweris readily understandable after a magnumatroxque
bellum.4 Tiberius also adopteda measurewhich was otherwiseseldomused:in
I I B.C. he had Pannonianyouths sold as slaves.s North of the Savevalley such
ruthlessmeasureswere unnecessary.In the Drave valley and in north Pannonia
potentiallydangeroustribal alliancesandconcentrationsof powerwereprobably
dissolved; as a result small tribes were able to gain their independenceunder
Roman control. Only in the caseof the Arabiatesand Hercuniateswere com-
pletely new groupsformed; both 'tribes',however,lived in boundaryzones;the
Arabiatesprobablybetweenthe spheresof influenceof the Boii andthe Taurisci,
and the Hercuniatesbetweenthoseof the (Boian) Celts and Pannonians.As a
result the Boian empire,alreadyweakenedby the Dacians,was dissolved;living
in its former territory were to be found Boii, Azali, Eravisci, Arabiates and
Hercuniates.
Thereis additional evidencein respectof the Azali. In ChapterI it was seen
that the Celticizationof north Pannoniahad alreadybegunin the fourth century
B.C.; substantialsettlementof Celtsin the Danubevalley is attestedby numerous
finds of the La Tene Band C periods. In the area occupiedby the Azali the
personalnamesof the indigenouspopulation in the imperial period are very
closely relatedto those of southernPannoniaand northern Dalmatia. Azalian
names,for which thereare generalanalogies,occurin the sameform in the Save
valley.6 The assumptionthat it was in Roman times that the Azali were first
transferredfrom south Pannoniato the Danubeis supportedby a recent epi-
graphicfind in Bulgaria. A tombstoneset up at the latestunderClaudiusfor an
Azalus statesthat he had servedin a cohort of freedmenformed by Augustus.7
But when and how was a north PannonianenslavedunderAugustus?The only
possibleoccasionseemsto be the suppressionof the Pannoniansin the Save
valley in I I B.C., when Tiberius sold their youths in Italy. It is thereforequite
possiblethat the Azali were still living in the Savevalley in I I B.C. and that it
was only by Tiberius that they were transferredto the Danube,where they
were given an areain the deserta Boiorum. It is probably not too bold to assert
that by this move Tiberius wished to smashBoian power. Thus in the middle
of the area dominatedby the north PannonianCelts a non-Celtic tribe was
settled.
It is not surprising that Tiberius adoptedsuch radical measureswhen re-
H
organizing local balancesof power. Roman initiative was behind the eastern
migration of the Marcomanni,the settling of the Iazygeson the plain and the
transferof enormousmassesof peopleto Moesia under Augustus.The proba-
bility of even more ruthless action is obvious in an area which was to be
constituteda province, particularly oneperquamvicinum to Italy. 8
Drastic measureswere necessaryonly in south Pannonia.Thereis no indica-
tion that Tiberiuseitherhador wishedto takemilitary actionnorth of the Drave.
The military occupation of the north was perhaps based on treaties, since
Carnuntum,for example,which had becomea legionary fortress by A.D. 15 at
the latest,andpossiblya little earlier,is attestedat the time of the outbreakof the
large-scalerevolt in A.D. 6 as 'a placein the kingdom of Noricum'9 with which
Rome had an alliance. Thereis no information concerningradical intervention
or military action by the Romans east of Carnuntum.The settlementof the
Azali was probablymadepossibleby the fact that the land allotted to them was
desertedafter the Boian-Dacianwar. In north-easternPannoniathe Eravisci
actually minted their own coins under Augustus on the Roman denarius
standard.Without somekind of agreementby the Romansthis would havebeen
unthinkable. In the main the silver coins of the Eravisci were an imitation of
republican denarii of the period 90-70 B.C., which had enjoyed very wide
circulation in Pannoniasince 35 B.C. The legendson theseimitations are some-
times abbreviatedpersonalnames(DOMISA, ANSA, DVTEVTA, etc.) but mostoften
the tribal namein the form RAVIZ or IRAVSCI (PI. 2). They were probably minted
in Aquincum, since their distribution occurs within a fairly narrow radius of
Budapest.Only a few reachedGaul and free Dacia. It is possibleto date the
coins, as it was not until the end of the first century B.C. that Roman money
beganto circulatein north-eastPannonia.In the hoardsof RAVIZ coins Augus-
tan denarii appear,andonehoardevencontainsa Caligulandenarius.I°
No doubt it can only have beenon the basis of a specialtreaty that a people
within the imperial boundariesestablishedby Augustus was allowed to mint
coins accordingto the Romanstandard,though with the imprint of their own
names. After the province of Illyricum was established,groups among the
indigenouspopulationwere not all treatedalike. Even after Bato had surren-
deredon the river Bathinusin A.D. 8, Tiberius appointedhim king underRoman
overlordship,and Bato set aboutestablishinghis political powerin his new role
in a ratherself-willed andindependentmanner.He was,of course,soondeposed,
but not at Tiberius' instigation. The peoplesin north Pannoniaagainstwhom
military actionwas unnecessary hadprobablyacknowledgedRomancontrol in a
variety of ways; in somecasesthe requirementsexactedfrom themprobablydid
not differ from the diplomatic arrangementsmade, for example, by Vinicius
56
Nativepopulationand settlemmt
with the Cotini, Osi and Taurisci outside the empire. It is not known which
tribes formed civitates liberae, foederataeor stipendiariae; it is, however,probable
that the treatmentmetedout to the Pannoniantribes in the Savevalley was the
most unfavourable,whereasthosein the north were dealt with moreleniently.II
In north Pannoniasporadicinstancesof citizenship grantedto natives are at-
testedas early as the first half of the first century. For example,C(aius) I(ulius)
Magimarusfrom nearAquincum was grantedcitizenshipeither by Augustusor
Caligula.I2 The fact that the north Pannonianswere not obliged to furnish
auxiliaries until the middle of the first century is perhapsalso attributableto
their more lenient treatmentenshrinedin a treaty. It was probably only in the
course of the first century that the differencesbetweencivitatesfoederataeand
stipmdiariaefaded into insignificance.
North Pannoniawas also usedin the Julio-Claudianperiodfor the transferof
peoplesfrom barbarianareas.Thereis literary evidenceof this only in the caseof
Vannius, king of the Quadi, who, togetherwith his followers, was settled in
Pannoniain 50. The early Romanmaterialexcavatedin Pannonia,mainly in the
areabetweenLake Pelso(Balaton) and the Danube,consistsof gravesand scat-
tered finds of a Dacian and early Germaniccharacter;this suggeststhat there
had beentransfersof peoplesfrom the areanorth of the Danube.I3 We should
probably not regard all thesefinds as the property of Vannius' people. The
Dacianfinds in imperial Pannoniago back to the Dacianizationof the northern
Carpathianregion, which beganin the middle of the first century B.C. and so
may also be evidenceof non-Daciansettlements.Particularly relevant in this
respectare the Cotini who were given land in Pannoniabeforethe beginningof
the secondcentury.In Ptolemy'slists they appearas Kytnoi in the north-eastern
part of PannoniaSuperior.I 4 The place-nameOsonibusin north-eastPannonia
could well suggestthe inferencethat the Osi, neighboursof the Cotini, were
settled there.Is Most of the Cotini and Osi, however, remainedbeyond the
frontier until the time of Marcus Aurelius. By the time he incorporatedthe
Cotini into the empire in the courseof his great war, the majority bore Dacian
names.I6 Even after the time of Marcus the Osi continuedto live outsidethe
empire.I7 The Cotini were basicallya Celtic peoplewho in the first centuryA.D.
were not yet completely Dacianized.In his Germania Tacitus says they spoke
Celtic.IS Thus the Cotini, when settledin Pannonia,were linguistically related
to thelocal population.It is possiblethat this wasalsotrue of Vannius' followers,
as the name Vannius is frequently attestedin Noricum and west Pannonia
among the original Celtic inhabitantsand must, therefore,be a Celtic (Boian)
name.I9 Onegroup of the Boii is also mentionedby Tacitus20 as living in south-
west Slovakia, that is, beyondthe frontier. It is not improbablethat Vannius,
57
king of the Quadi, was descendedfrom a noble Boian family and that his
followers consistedof the Boians outside the empire who had come under
German rule. A significant proportion of the groups which were settled in
northernPannoniain the courseof the first centurywerethus probablyof Celtic
origin.
Among them was a womanbelongingto the Celtic tribe of the Anartii, who
had settledat Aquincum in the first half of the first centuryand had beengiven
Roman citizenship.21The settlementsof Celts and Germansfrom outside the
empire were dispersed over a fairly extensive area and their communities
attachedto the existingadministrativeunits (civitates).Thereis no indicationthat
they formed their own civitatesperegrinae.In Ptolemy'slists the Cotini alone are
mentioned as having a continuous settlement area; however, his map of
Pannoniais so distortedthat it is impossibleto locatethemaccurately.The most
likely area would be south of Lake Pelso, which in Romantimes was heavily
wooded and in part marshy. According to Ptolemy the Kytnoi must have
inhabitedthe areabetweenthe Azali and the Iasi.
On the basisof indigenousfamily names,inscriptions,local costumesdepicted
on tombstonesand the jewelry which went with thesecostumes,it is possible
to give a fairly accurateoutline of the ethnic scenein Pannoniain Romantimes.
The sources-stonemonumentsand grave-goods-althoughthey begin only
towardsthe end of the first century,permit of inferenceswhich are valid for its
first half. It will probably be easierto understandthe later periods if the in-
formation derivedfrom therelevantsource-materialis summarizedat this point.
Largish closedgroupsof indigenous(i.e. non-Latin) names22 are found only
in some areasof the province. In other parts they occur only sporadicallyin
inscriptions. There are also large areas in which no inscriptions have been
found, but in many caseswe can supplementtheir evidencewith that of military
diplomas which were issuedto membersof Pannoniantribes, and which bore
the nameof the auxiliary soldier and of the tribe to which he belonged.Since
nameson military diplomascan be classedin the sameonomasticgroup as those
on inscriptionsin the areain question,the diplomascan also be usedas sources
for the classificationof family names.Non-Latin cognominain the three-name
systemare taken to be indigenousnames,provided it can be proved that the
bearerof the Romanname was not a foreigner, as also are the single namesof
peregriniin so far as it can be provedthat they were neitherslavesnor foreign
auxiliaries.
Peregrini are attestedin closed groups in the upper Save valley, in north-
west Pannonia(particularly in the region of the NeusiedlerSee, Lake Ferta)
and also in the north-eastof the province. A well-defined group was found
58
Nativepopulationand settlemmt
in the upper Save valley, south-eastof Emona. Within this group there are
nameswhich otherwise do not occur in Pannoniabut in contrastare all the
more frequent in Istria and north-eastItaly, where they form the northern
Adriatic group of Venetic-Liburnianfamily names.Among the tribes in Pan-
nonia, though without indicating their settlementarea, Pliny lists the Catari;
they were probably relatedto the Catali who dwelt aroundTergeste(Trieste).
The Venetic-Liburniangroup of namesin the Emonaregion can thereforebe
regardedin the samelight as thoseof the Catari. Elsewherein the upper Save
valley the namesare clearly Celtic; there are analogiesmainly in neighbouring
Noricum. Names in north-west Pannonia are also Celtic, again with their
closestanalogiesin Noricum. Thoseof the Leitha area,however,are somewhat
influenced by Germannames,such as Strubilo, Tudrus, etc., which are often
thought to reflect the settlementof peoplefrom outsidethe empire and, more
specifically, that of the followers of Vannius. Those with indisputably Ger-
man names were, however, freedmenof the original inhabitants with Celtic
names;they must,therefore,havecomeinto the provincevia the slavetradeand
not as the result of the settlementpolicy describedabove. In north-eastPan-
nonia indigenousnamesfall into two distinct groups.In the westernpart of the
region wherethey occur somehaveeither no analogiesor are attestedin south-
ern Pannonia,particularly in the region of Sirmium and amongthe Breuci (e.g.
Bato, Breucus, Dases,Dasmenus,Licco, Liccaius, etc.). Although there are
Celtic names(e.g. Aturo, Busturo,Ciliunus and Saco)on the easternedgeof this
area, theseare unknown among the rich variety of their easternneighbours'
names.The westerngroup of names,which can certainly be identified with the
Azali and the easterngroup of the Eravisci, is contiguouswith a north-eastto
south-westline from the bend on the Danubeto Lake Balaton, which formed
the boundarybetweenPannoniaSuperiorand Inferior. According to Ptolemy
the Azali dwelt in Superior and the Eravisci in Inferior. The Azalian names
suggestthat a southPannonianIllyrian peoplewas superimposedon a numeric-
ally weak Celtic population.The Eraviscannamesare most closelyrelatedto the
Boian in the Leitha areaand to Celtic namesgenerallyin Noricum. Non-Celtic
elementsare found only sporadicallyand peripherallyamongthe Eravisci, e.g.
Bato in the south of the tribal area. Otherwisethe namesof the Eravisci were
not only undoubtedlyCeltic but equally undoubtedlybelongedto the Norican-
Boian stock. Two passagesin Tacitus' Germania seem to rebut this. In de-
scribingthe barbarianareasof Germanyhe saysthat the Cotini andOsi who lived
betweenthe Germansand Sarmatianswere neitherGermannor Sarmatian;the
Cotini spokeCeltic andthe Osi 'Pannonian'. 23 Otherwise,thereis no mentionin
ancientliteratureof a 'Pannonianlanguage'.Earlier researcherstook the lingua
59
Native population and settlement
Pannonicato be Illyrian, and sincein anotherchapterof the Germania24 Tacitus
lists the Eravisci and Osi as being linguistically relatedtribes living on either
side of the Danube,it was inferred that the Eravisci were Illyrians. Recentre-
searchersin the analysisof indigenousnameshave now provedthat the names
of the Eravisci were completely Celticized; the use of 'compound'and 'de-
scriptive' names(e.g. nameswith the suffix -rix = king) indicates an under-
standingof CeltiC.25 Thus thereis no doubt that the Eravisci spokeCeltic. What
then was the 'Pannonianlanguage'mentionedby Tacitus? He supplies the
answerhimself when explaining his hypothesisconcerningmigrations to and
fro acrossthe Rhine and the Danube.He assumesthat theserivers are linguistic
boundaries,i.e. that the Rhine separatedthe Germansand Celts; accordingto
him Celtic tribes east of the Rhine and Germantribes west of it were settlers
who had migratedacrossthe river. A linguistic relationshipbetweentribes on
eitherside of the Rhine andDanubehadthereforeto be explainedby migrations
acrosstheserivers. Sincethe Osi andthe Eraviscilived on the oppositebanksof
the Danubeand spokethe samelanguage,either the Osi must have migrated
from the Eraviscanareaor vice versa.Tacitusis, of course,unableto provewhat
happened,althoughit was he who propoundedthe whole migrationhypothesis;
so he admits that the direction takenby the migration is uncertain.In his main
source,the Elder Pliny's history,he had only found the statementthat the Osi
and Eravisci spoke the samelanguage,and as he did not regard Pannoniaas
belongingeither to the 'Gallic' (i.e. Celtic) or to the Germanlinguistic areahe
was forced to the assumptionthat there was a separatePannonianlanguage
which he then ascribedto the Osi.26
Thesepassagesin Tacitus have called for closerexaminationsincehis theory
of a migrationby the Osi or the Eravisci hasplayeda key role in the reconstruc-
tion of the history of the first centuryB.C.27 The only grain of truth in Tacitus'
hypothesisto survive is that the Osi, Cotini and Eravisci belongedto one and
the sameCeltic group of peoplesin the northernCarpathianregion. It was not
until the first centuryA.D. that fate intervenedto separatethem; the Cotini and
Osi outsidethe empirecameunderthe control of the Quadiand Sarmatiansand
were later gradually Dacianized,particularly after the large-scaleflight from
Dacia in the time of Trajan, whereasthe Eravisci were able to maintain their
Celtic languageand culture under Roman rule. At one time, however, these
peopleswere in the 'samestateof misery' accordingto Tacitus.
As for the areaswhereindigenousnamesoccuronly sporadicallyor in military
diplomas,it would seemthat thoseof the Varciani were still Celtic, but thoseof
the ColapianialreadyIllyrian 28 (Pannonian).Thoseof the Iasi, Breuci, Amantini
and Scordisciare south Pannonian-Dalmatian. It is thereforehighly probable
60
that the namesof the Oseriateswho lived in the Savevalley betweenthe Breuci
and Colapiani also belongedto the south Pannoniangroup. The sameassump-
tion may be madefor the Andizetes,in view of the fact that Strabo regarded
them as Pannonians.Celtic nameswith a Norican characterare also met with in
the Dravevalley west of the Iasi, and occasionalnamesnorth of the upperDrave
as far as the NeusiedlerSeeare likewise Celtic-Norican. SporadicCeltic names
are also found to the south of the Eravisci.
There are of coursegaps in this distribution of namesin Pannonia,as for
some sizeableareasthere is no available material. Nevertheless,a fairly clear
picture of the linguistic groupingsof the indigenouspopulationdoes emerge.
Celtic tribes inhabited the western and northern parts of the province, and
Pannonian(Illyrian) tribes the remainderas well as a wedge-shaped areawithin
the Celtic populationin north Pannonia(Fig. I I, p. 64). We have seenthat this
southPannonianlinguistic island in northernPannoniacan be tracedback to a
settlementorderedby Tiberius. In the first centuryB.C. Celts were still living in
the areaof the Azali; it was only possiblefor a south Pannonianpeopleto take
it over after the Daciansunder Burebistahad laid waste the land of the Boii.
Quite a small areaaround Emona belongedto the Venetic-Liburniantribe of
the Catari.
The stock of Celtic namesis strikingly uniform; from this it may be inferred
that the Celticization of west and north Pannoniawas both protractedand in-
tensive. Contactswith the Celtic inhabitantsof Noricum were of the closest.A
surprisinglylarge numberof Celtic namesin Pannoniaalso occur in Noricum.
Celtic namesattestedin Gaul or in other western Celtic areasare much less
frequent.29 Among them there is, however, a small group which indicates
contact with Spain (Aturo, Ciliunus, Teitia, Anbo, etc.). This detail could be
ignored were it not for the fact that these namesoccur in the north-eastern
corner of Pannonia,where a very striking type of tombstonefrom the early
imperial period revealsthe closestpossiblecontactwith Spain.30 Thesetough,
very simply executedtombstoneswere set up for indigenousperegrini;they are
characterizedby the so-called'astralsymbols'(Fig. 10) suchas the sickle moon,
sun disc (often depictedas a rosette),'gatesof heaven',starsand so-called'keys
of heaven',etc., all of them symbols which occur on tombstonesin Salamanca.
How thesecharacteristicallySpanishinfluencesfound their way into tombstone
symbolismandaffectedpersonalnamesin Pannoniacannotas yet be determined.
They do, however,point to the possibility that Celtic culture in Pannoniadoes
not derive solely from the Norican-Alpine Celtic province. Researchis faced
with the task of separatingout the different componentsof this Celtic culture
in Pannonia.It has only recently beenestablishedthat Celts from the Herrynia
61
NORTH~EAST
PANNONIA
Figure IO Tombstoneswith astral symbols
62
silva also went to Pannonia(Latobican house-urns;cf. also the tribal name
Hercuniates).3 1 A passage in Caesarmay be quotedin supportof the southern
migration of the Latobici.32 The very confusedlegendof the return and settle-
ment of the Tectosagesin Pannoniacan perhapsbe consideredin connection
with the Spanishcontacts.33
The ethnographiccharacteristicsillustrated by local costume and jewelry
suggestthat therewas an extensiveNorican-Pannonian cultural provinceof the
Celts. This province, as far as women's clothing was concerned,was divided
into several smaller costume areas (Norican hood, turban-veil, apron, etc.)
which can unfortunately only be attestedwhere women are depictedby full-
lengthfigures, or at leastby busts;jewelry, however, whichis the sameboth in
gravestonesand in finds is distributedover the sameareaswhere Celtic names
occur. This jewelry consistsin the main of the so-calledNorican-Pannonian
wing-brooches (Fliigeljibeln) (PI. 4a), a characteristic belt-fitting frequently
ornamentedwith open work, and a variety of other pieces of jewelry, e.g. a
richly ornamentedneck-charmand pair of brooches,attestedon gravestones
but only recently found in graves (the so-called bow-tie shaped brooches
(Maschenftbeln)). These pieces of jewelry are restricted to the western and
northernparts of Pannonia.They do not occur amongthe genuinePannonians
in the south.34
Information concerning ethnic and linguistic divisions among Moesia's
indigenouspopulation is very sparse(Fig. II). The reasonfor this is not so
much the lack of investigation but rather that neither indigenous names,
costumesnor anything else which could lead to a grouping of the original
inhabitantshas come to light. There is only sporadicepigraphicevidenceof
names,and, to make mattersmore difficult, it is not always clear whetherthe
non-Latin nameson the inscriptionsare those of native local inhabitantsor of
nativesof someotherBalkan area.For example,at the endof the secondcentury
A.D. in the mining areaof Mount Kosmaj, south of Belgrade,thereis mention
of a considerablenumber of peregrini with Illyrian and Thracian names. It
would appear,however,that they went there to work in the mines when these
were startedunderMarcusAurelius,35The only known group of local peregrini
is that in the Metohija in the Dardanianarea.36 There, too, the many busts on
tombstonesareto be regardedas depictinglocal costume,but the very primitive
quality of the stonemason's work permitsonly of very generalinferences,as,for
example,that broocheswere probablyunknownand that a little cap was worn,
not unlike the felt caps popular even today in the central Balkans.Women are
mainly depictedwearing a veil which hung straight down from the headand
which likewise has persistedto the presentday.37 Thesegeneral observations
63
an
CL
aci
CL
Thr
CL
CL
n
acia
Thr
A ARABIATES
CL COLAPIANI
CR CORNACATES
H HERCUNIATES
o OSERIATES
o, 100 200
V VARCIANI km.J
Languages:
Celtic ll1yrian Venetie Thracian
(Pannonian)
Figure I I Languagesand tribal nameswith Latin suffixes
64
are, however,not of muchhelp, sincethereis a lack of materialfrom otherparts
of the provinceon which to basea comparison.Apart from thesethe analysisof
namesis basedon thoseof legionariesfrom Scupi and Ratiaria, someof whom
hadindigenousnames.From this scantymaterialthe following outline emerges.
Names,where certainly local inhabitantsare concerned,are clearly Thracian
with a south Pannonian-Dalmatian (Illyrian) colouring which is found almost
everywherebut is not completely characteristic.The list of namesof the 196
dischargedveteransoflegio VII Claudia38 includes,for example,soldiersfrom
Scupi and Ratiaria called Dassius.Perhapsthe nameAndio, variants of which
(Andinus,Andia, etc.) are attested in Dardania,shouldalso be added.The rest of
the soldiersfrom ScupiandRatiariaon this list havegoodThraciannames,such
as Bitus, Sinna,Dolens,Drigissa,Mucco, Auluzon, Mucatral and Daizo. Apart
from Illyrian influence, there is also evidenceof southerncontact; the name
Mestrius, the most frequent amongthe natives of Scupi, is just as commonin
northernMacedoniaamongPaioniansand Pelagonians.With both Dassiusand
Mestrius (Mestrianus)it is a questionof single names.Thus it is possiblethat
popular nameswere taken over by neighbouringlinguistic groups.
The Illyrian elementis strongeramongthe Dardaniansin the Metohija.39 In
addition to the commonly occurring Das(s)ius, the names Scerviaedusand
Andia may both be regardedas Illyrian. Otherwise,namesin this group are
either Thracian(Sita, Nanea)or uniquely occurringnames(Cittu, Timens,etc.).
This material is too meagreto provide an answerto the old questionwhether
the Dardanianswere Thraciansor Illyrians. However,a generalconclusionmay
be permitted,that the original inhabitantsof MoesiaSuperiorwere in the main
Thracian, but had been exposedto Illyrian influence from the west, with the
result that the Dardanianareain particularemergesas the contactzonebetween
the Illyrian and Thracianlanguages.The inhabitantsof Scupi probably spoke
Thracian,as a Romansoldier born therein the third centuryconsideredhimself
a Bessus.40 In late antiquity Bessuswas the normal term applied to Thracian-
speakinginhabitantsof the empire;the lingua Bessicawas Thracian.Thracianand
Illyrian namesoccurin one and the sameDardanianfamily in the Metohija, e.g.
ScerviaedusSitaes(Sitae filius) 41 (PI. 5a). Since,however,it lies on the border
betweenDalmatiaand Moesiathe Metohija cannotbe regardedas the yardstick
in determininglinguistic classificationthroughoutthe whole ofMoesiaSuperior.
Henceit is betterto assumethat the indigenousinhabitantsof that area spoke
Thracian.
Namesshow hardly any evidenceof the Celtic elementrepresentedin Moesia
Superior by the Scordisci in particular. Some namesin the westernMorava
valley in the Dalmatian-Moesianborderareawerein fact takento showtracesof
65
Celtic influence, but this is anything but certain.42Strabo and other sources
mentionan Illyricization and Thracianizationof the Scordisci,dependingupon
which racehadsubduedthe latter.43 It has beenestablishedfor Pannoniathat the
namesof the ScordisciaresouthPannonian-north Dalmatian,i.e. they havebeen
Illyricized; the Thracianizationobviously refer~ to those Scordisci who had
managedto retain their hold on the northernpart of MoesiaSuperior.
Evidenceof the settlementof barbariansin Moesiais providedby two written
sources.Aelius Catus in the reign of Augustus settled 50,000 Getae,44and
Plautius Silvanus Aelianus in that of Nero had more than 100,000 'trans-
Danubians'settledin Moesia.45 Thereis much to supportthe assumptionthat
part at least of these settlementstook place in the area which later became
Moesia Superior. In all probability those settledin that area were Getae,and
were linguistically relatedto the indigenouspopulationof Moesia. Finally, the
Daciansin Moesia must be considered.A diploma of 71 was issued to one
Dacus,46which suggeststhat therewas a Daciancivitas on the right bank of the
Danube.This will, however,be dealtwith later.
After the constitutionof the provincesthe native populationwas assignedto
administrative units known in most of the imperial provinces as civitates
peregrinae. Thesecivitates were either called after the local tribes and peoples
(civitas Boiorum, Azalorum,etc.) or were given namesderivedfrom the organiza-
tional activity of the Romangovernment(civitas Colapianorum,Cornacatium,etc.,
seep. 53). The territories of the civitates did not necessarilycorrespondin size
to thoseareasin the handsof a given tribe in pre-Romantimes. As has already
beenpointed out, the pre-Romantribal areaof the Breuci was probably much
bigger than the territory of the civitas Breucorum. Organizing the civitates
probablytook quite a long time. The radical measuresadoptedby Tiberius after
the conquestof Pannoniaclearly requiredseveralyears,as did thosetaken after
the rebellion had been suppressed.It is very possiblethat the administrative
framework was not finally completeduntil after A.D. 8.
The civitatesperegrinaementionedby Pliny andPtolemy,in largely correspond-
ing lists, were,moving down-streamalongthe Danube,thoseof the Boii, Azali,
Eravisci, Hercuniates,Andizetes, Cornacates,Amantini and Scordisci; down-
streamon the Save,thoseof the Catari,Latobici, Varciani, Colapiani, Oseriates,
Breuci, Amantini and Scordisci; down-streamon the Drave, those of the
Serretes,Serapilli, lasi, Andizetesand, finally, betweenthe Serapilli and Boii,
the Arabiates.The Belgites, mentionedonly by Pliny, cannotbe located.Only
laterin the courseof the first centuryA.D. wasthe civitas Cotinorumorganized-
probablyto the southof Lake Pelso.
Thereis a markeddiscrepancybetweenthe lists of Moesiantribes given by
66
TRICORNIUM
TRIC .PINCUM
TRIC
ORN
ORN
TRICIUM
IU
TRIC M
TRIC
TM
ORN
RICO
TR
TR
IC
TO
ORN
TRIC
OR
RR
IU
ICOIR
RNIU
ICNOIU
NI
UM
ORN
UM
RM
NIUM
M
NIUM
IUM
TRICORNIUM
TRICORNIUM
TRICORNIUM
land over 1500 metres

o 50 100/~
, I ",m
Figure I 2 The native peoplesof Upper Moesia

Pliny and those given by Ptolemy.47This suggeststhat there was probably a
reorganizationof the civitatesperegrinaein Moesiain the first century. Pliny lists
the Moesiantribes as follows: Dardanians,Celegeri,Triballi, Timachi, Moesi,
Thraces Pontoque contermini Scythae. The sequenceis geographical,not alpha-
betical, and henceis not basedon the material collectedby Agrippa for the map
of the empire. Since Moesia was constituteda province only after Augustus,
Pliny's list probably datesto the time of Tiberius. In supportof this dating is
the fact that eastof the Triballi, in the areawhich was laterMoesiaInferior, only
the territory of the regnum Thraciae and the Scythiansare mentionedin general.
The Celegeriand Timachi are attestedonly by Pliny; they lived respectivelyto
the north of the Dardanianson the western edge of the province and in the
Timok valley. There is satisfactoryproof from other sourcesthat the Moesi
lived around Ratiaria.48 In contrast, Ptolemy lists the following tribes: the
Tricornensii on the Dalmatian boundary, the Moesi on the river Ciabrus
(=Cibrica), and betweenthem the Picensii and the Dardaniansin the neigh-
bourhoodof Macedonia.This list probably correspondsto the situationin the
first decadeof the secondcentury. It would seemthat at some time between
Tiberius and Trajan a reorganizationtook place which was perhapsconnected
with the establishmentof the Roman administration of what later became
MoesiaInferior.This reorganizationinvolved replacingthe CelegeriandTimachi
by the TricornensesandPicenses,tribes calledafter place-nameson the Danube.
Tricornium is the ancient name of Ritopek, east of Belgradeon the Danube,
while Pincumis the ancientnameof Gradisteat the mouthof the Pek( = Pincus).
The native communitieseastof the Triballi listed by Pliny also took their names
from placeson the Danube,e.g. Dimenses,Oetenses,Obulensesand Appiaren-
ses.This reorganizationin MoesiaSuperiorandInferior thereforeprobablytook
place before the division of that province in 86. The actual timing was either
c. 46, when Tullius Geminus,the governor,was ordering territorial affairs, or
during the governorshipof Plautius Silvanus Aelianus, who under Nero was
settlinglarge numbersof trans-Danubians in Moesia.In any caseit is significant
that the administrativeframework in Moesia was establishedonly gradually by
the imperial government,in stages; interferencein the affairs of the native
population was neither so radical nor so abrupt as in Pannonia.There were
probably also short-lived communities, such as that of the Dacians, which
neitherPliny nor Ptolemy mention. However, accordingto a diploma dated71
they also must have formed a civitas peregrina in the secondhalf of the first
century.49
The renamingand reorganizationwere obviously connectedwith the settle-
ment of large numbers of people from the left bank of the Danube. From
68
Tiberius until the second century perhaps only the Dardanians,Moesi and
Triballi escapedthesereforms.Theseracesarethe only onesmentionedby both
Pliny and Ptolemy. In the early decadesof the first century theywere already
organizedinto communitiesunderRomancontrol; in fact they inhabitedareas
which were the first to be placed under military occupation.The fortress of
legio IV Scythicamusthavebeensomewherein Dardania;that of V Macedonica
was transferredat a very early date to Oescuson the Danubeand a numberof
auxiliaries were stationedin the vicinity. 50 Ptolemy refers to Dardania as a
specialdistrict in Moesia Superior.5! It is very probablethat in the last yearsof
Augustus,even before the province was constituted,the Dardaniansformed a
communityunderRomancontrol. Thereis evidencethat this was so also in the
caseof the Moesi, and Triballi, who were organizedas civitates roughly in the
seconddecadeof the first century A.D. under the names Moesia et Treballia.
Supremecontrol over thesecivitates was in the handsof the senior centurions
belongingto the legion stationedin Oescus.In this capacityC. BaebiusAtticus
is epigraphically attestedas being primipilus leg. V. Mac. praeJectuscivitatium
Moesiae et Treballiae.52 In Dardaniathe primipilus of nn Scythica may well
have beenpraeJectusof the civitas Dardanorum.
The administration of the civitates peregrinae in Pannoniaand Moesia was
probablyestablishedalong the lines of theseearlypraeJecturae.Thereis evidence
of a praeJecturaof the civitas Colapianorumunder Claudius: AntoniusNaso, as
primipilus of legio Xln Geminastationedat Poetovio,was entrustedwith it.53
In the early Flavian period L. Volcacius Primus,prefectof cohorsI Noricorum
stationedat that time at Arrabona,was simultaneouslypraejectusripae Danuvii et
civitatium duarum Boiorum et Azaliorum.54 This military control over the civitates
peregrinae outlasted the Julio-Claudian period. How the praeJecturae were
allocatedbetweenofficers of Equestrianrank and primipili thereis no meansof
telling. The primipilus probably did not hold the praeJecturaover all the neigh-
bouring civitates; evenAntonius Naso waspraeJectusonly of the Colapianiand
not of all the other civitates in south-westPannonia.The Boii who lived in the
neighbourhoodof legio XV Apollinaris hadVolcaciusPrimus,a cohort prefect,
as their praeJectuscivitatis, whereasthe primipilus of that legion was probably
chargedwith the supervisionof anothercivitas. The civitates of easternPan-
nonia, which were at a considerabledistancefrom the legionary fortressesat
Carnuntumand Poetovio,probablyhad aspraeJectusthe commandingofficer of
an auxiliary unit stationedeither within the areaor nearthe civitas in question.
The Julio-Claudianforts, mentionedin the previouschapter,were also probably
the headquartersof military administration.It is not known whether all the
civitates came under this military administration. Some of them, which had
69
offered no resistanceunder Augustus and hencehad perhapsbecomecivitates
liberae orJoederatae,probablyhadno military praefecturae.55 Thelate establishment
of auxiliary forts on the Danubepossiblysupportsthis hypothesis.A definitive
statementis, of course,impossible.
The officers and their soldiers could not carry out all the tasks involved in
administeringthe communities.Even the military were dependentupon a local
body or at leastsomerepresentativesof the population.The probability is that
the local social system, its class structure and its leading bodies were not
abolishedout of hand. Drastic interferencein the social systemmay perhapsbe
inferredfrom an epitaphto a young man of the Amantini, who died while held
hostagein Emona, probably under Augustus. It is emphasizedthat he was
Amantinusgente Undius centuria secunda.56 The division of the tribal clans into
centuriaecould have beenbasedon a Romanenactment,wherebythe Amantini
who were engagingin active resistancewere organizedinto centuriaeas a means
of more effectively controlling them. It is, on the other hand, just as likely that
the hostagesdetainedin Emona were organizedinto centuriae.57Inscriptions
from the secondhalf of the first century and from the first half of the second
century contain referencesto native principes. The earliest to be attestedepi-
graphically is perhapsIucundusprinc. Azalus.58 Theseprincipes were members
of a body on which Romancontrol could rely. The criteria for the selectionof
theseleading men probably did not differ from those applied universally by
the Romangovernment.The sameformula is usedby Claudius,Aelius Aristides,
Antoninus Pius and others:'the best(i.e. the most loyal) andthe richest'.59 The
local aristocracy, whose members were gradually given a share in Roman
control asprincipes,and adaptedtheir intereststo thoseof the Romans,acquired
Roman citizenship in increasing numbers-someof those attested already
possessed it 6o-and by the secondhalf of the first century alreadyformed the
most importantsupportfor Romancontrol. A very good yardstickin assessing
the consolidationof this control, as has alreadybeenseen,is the fact that there
was a gradualrelaxationof military occupationin the interior of the province;
on the one hand the total number of troops was reduced(Claudius actually
withdrew a legion) while on the other the centre of gravity was gradually
shifted to the Danubefrontier.
From the Julio-Claudianperiod there is only very sparseevidencethat the
local aristocracyin the civitates copied the Roman way of life and acquired
Romanindustrial articles.The reasonwas probablythat this classwas numeric-
ally weak at that time. So far neither cemeteriesnor settlementscontainingim-
ported Roman goods from the pre-Flavianperiod have been uncovered.The
cemeteriesin the legionary fortressesand colonies are, of course, the excep-
70
tion. 61 In graveswhich are clearly thoseof natives,Romanpottery, glassesand
bronzeobjectshavebeenfound only on the westernedgeof the provincealong
the amberroad, andparticularlyin the tumuli betweenthe Drave and Danube.62
From this we may infer that the native populationcould neither afford foreign
(Roman)goodsnor was interestedin acquiringthem. Romantradershad adap-
ted their tradeto the demandsof the military and foreign settlers,while Roman
craftsmenhad establishedtheir workshopsonly wheretherewas a call for their
wares.Along the amberroad, from Aquileia to Carnuntum,trade was already
developingin the Julio-Claudian period; it was thus possible for the native
populationliving alongthis road to buy importedRomangoods.The marketto
which the north Italian tradersdirectedtheir attentionwas, however,createdby
foreigners.
Foreigners,in the main north Italian newcomers,were alreadybeginningto
appearin PannoniaunderAugustus.According to Velleius, on the outbreakof
the rebellionof A.D. 6 'Romantraderswere massacred'. 63 It is, however,not out
of the questionthat his statementis no more than an indispensableingredient
of any descriptionof an anti-Romanrebellion. Somewhatlater thereare reports
of Roman traders settling permanentlyin the Marcomannicempire;64 in fact
cemeteriesof early imperial datein Bohemiahaveproduceda very rich store of
imported Roman goods, particularly bronze vessels,altogether a very much
richer yield than that in Pannoniangravesof the sameperiod.65 Rome'strade
with the Marcomanni-inhis mention of it Tacitus refers to the ius commercii-
probablystartedafter the legionaryfortressat Carnuntumhad beenestablished.
Along the amber road,at Emona, Poetovio and Carnuntum,inscriptions re-
ferring to Italian, mainly Aquileian, civilians date backas early as the first half of
the first century. Among them is a striking number of freedmen who, as
representativesof north Italian trading houses,managedbrancheswhich had
beenestablishedin Pannonia.66 An important part in developingthis network
of trade was played by the fortressesat Carnuntum,Poetovio and possibly
Siscia, wheretradersand craftsmenwho cateredfor the needsof the army had
settled. This group of small- and large-scaleentrepreneurs,at times only
toleratedand at others welcomed,led to settlementsaround the fortressesas
early as the first half of the first century; in the oldest cemeteriesat Poetovio67
and Carnuntum68 tombstonesset up to both legionariesand civilian traders-
mostlyfreedmen-werefound. We mustincludeherea tombstoneof a manfrom
Tergeste(Trieste) locatedat Siscia, where investigationshave not yet begun,69
and dating to the time of Augustus or Tiberius. There is copious epigraphic
evidence of branches of foreign (north Italian) trading houses in Emona,
Poetovioand Carnuntum.70
71
Native populationand settlement
Archaeologicalmaterialhas so far failed to provide sufficiently clear informa-
tion about the nature of this trade. The earliest products of Italian industry
in imperial times have been discoveredin only a few placesand mainly in the
form of loose finds; it is, however,no accidentthat Arretine sigillata (but only
laterforms) (Fig. 20), andbronzevesselsfrom Campania,Aco-beakersandnorth
Italian glass'-waredating from the Julio-Claudianperiod have beenfound only
in the settlementsalongthe Amber Roadand in a few placesin the valleys of the
Drave and Save.'! Thereis much to be said for the view that thesegoodswere
bought only by the military and by foreignerswho had settledin the province.
The indigenouspopulationwas still living under conditionswhich had existed
since time immemorial, and used only local productswhich cannot accurately
be described, since there is still insufficient information about the material
culture of the later pre-RomanIron Age. Onceagainit may be possibleto draw
an inferencefrom a later period: in the secondcentury there were still ceme-
teries and settlementsin which only local goods were found, and imported
goodsare either not presentat all or only on a very minor scale.In generalit is
right to assumethat the indigenouspopulation in the Julio-Claudian period
maintainedtheir late Iron Age mode of life intact.
At presentit is impossibleto decidewhetherthe only thoroughlyinvestigated
late La Tene settlementin Pannonia,that on Mount Gellert at Budapest,7 2 can
be used as a basis for assessingthe transitional period. This settlementon the

steep rocky hill in the centre of the Hungarian capital acquired importance
becausethe Danube below is very narrow at this point and routes from the
Carpathianbasinhaveto crosshere.As a consequence Budapestwas always the
most importantcrossing-placein the area.An oppidumwas establishedon Mount
Gellert in the first centuryB.C., and at its foot, nearthe ancientferry in the part
of the town known as Tabin (Fig. 14), an opensettlementwith a markedindus-
trial and commercialcharacterdeveloped.This settlementwas protectedby the
oppidum.The traditional red-and-whiteceramicsof the late La Teneperiod were
producedin the industrial centreat Tabin. The reinvestigationby E. Bonis of
the rich archaeologicalmaterial has revealedthat there was a wide ramification
of cultural and commercialcontacts.Thereis no evidenceof Romaninfluences,
although part of the material must be dated to the early decadesof Roman
control. Only the imprint of an intaglio, depictinga seatedVictoria with shield,
with which the red and white paintedvesselswere stamped,showsany Roman
influence(PI. 4b).
The Mount Gellert hill-fort was abandonedaroundthe turn of the first cen-
tury B.C.; the settlementof Tabin probably lasted somewhatlonger. After the
Roman conquest native fortified hill-settlements were either evacuatedor
72
gradually abandonedthroughout the whole of Pannonia. Excavations on a
smallerscalehaveprovedthis to be so in the BraunsbergnearCarnuntum,73the
LeopoldsbergnearVindobona74 and at others,too. Nor did the big oppidumof
GomolavanearHrtkovci,75 not far from Sirmium, produceany Romanmaterial
worth mentioning. Most earlier investigatorswere of the opinion that these
fortified oppida representeda dangerto Roman security and hencehad to be
evacuated.It is, however,not unlikely that theseuncomfortable,overcrowded
sites, which were apparentlyonly used as defencesin an emergency,simply
ceasedto havea purposeafter the Romanconquest.It is probablethat after the
abandonmentof the Mount Gellert oppiduma settlementdevelopedto the north
of it on the site of the later legionary fortress (Fig. 14), whencearchaeological
material mainly of the late La Tene type has beenexcavated.It is assumedthat
this settlementsprangup aroundthe auxiliary fort at Aquincum. Sucha fort is
in fact assumedto haveexistedtherefrom the middle of the first century.76It is
possiblethat the industrial quarterat Tabin was also transferredto the settle-
ment, where the military provided a market for its products.It was at Aquin-
cum, too, that the gradualtransformationof industry and the native way of life
receivedits first impulsefrom the army. At siteswhereno troopswere stationed
life went on as before.
What has just beensaid appliesprimarily to Pannonia.Investigationof Julio-
Claudianforts and native culture in Moesia Superiorhas still to be carried out.
Such loose finds of imported Roman goods as have so far come to light are,
however,of a remarkablylate date and belong to the Flavian period. Whereas
the settlementof Italian tradersin Pannoniacan also be attestedepigraphically
from the first half of the first century onwards,in Moesia Superiorthereare no
inscriptionseither to soldiersor to civilians of the Julio-Claudianperiod. This
absenceof inscriptionsis all the more important for judging conditions in the
country, sincefrom the time of Augustus,as in Pannonia,two legionshad been
stationedin Moesia. In the fortress of Viminacium inscriptions do not begin
until very much later, while owing to this absenceof inscriptions the other
fortress has so far not even beenlocated. We have already mentionedthat, as
early as the beginningof the first century, civilian settlementswere established
around Pannonianlegionary fortresses,and tombstoneswere set up by their
richer inhabitants.That such stonesare unknown in Moesia is possibly proof
that Italian tradersdid not arrive until muchlater.
The vicinities of fortresseswere not the only places where foreign settlers
establishedthemselves.A small group of auxiliary veteranssettled,for example,
north-west of Scarbantia,near Walbersdorf.77 Their tombstonessuggestthat
they were set up to severalinterrelatedfamilies who kept slavesand freedmen.
73
Pliny calls Scarbantiaoppidum ScarbantiaIulia,78 which points to some autono-
mousgroup of Romancitizensin the neighbourhoodof Scarbantia.Suchgroups
of citizens,occasionallyenjoying a kind of autonomy,settledmainly, of course,
nearlegionaryfortresses.In Scarbantiathe nameIulia suggeststhat autonomy
had been grantedby Tiberius. Similar groups were to be found in west Pan-
nonia, in the upper Savevalley and probably elsewhere,though of the latter so
far thereis no proof. In Scarbantiaone of the earliesttombstoneswas set up to a
dec(urio) Scarb(antiae) who probably was not yet a decurion of the Flavian
municipium(otherwisethe inscription would havereaddec. mun. Fl. Scarb.),but
a decurionof the oppidum.7 9 It is also not impossiblethat the citizens of Emona
who dedicatedan inscription in Savaria to Aequorna, a goddessfrom their
homeland, were already settled in this area in pre-Claudiantimes. On this
beautifully executedplaquethey refer to themselvesasconsistentes jinibus Savar(iae)
80
and notjinibus coloniae Claudiae Savariae. From the time of Claudius this latter
title was consistentlyusedon inscriptionsin the town.8I
In the Julio-Claudianperiod not a single town with Romanmunicipal status
was establishedin Upper Moesia, and there were only two in Pannonia.Both
were coloniae which were founded by settling legionarieswho had completed
their military service,and both were sited on the amberroad (Fig. 37, p. 220).
Emona (Ljubljana) (Fig. 13) was founded by Tiberius, but settlementof
veteransbeganin the last years of AugustuS.82 The connectionbetweenthe
mutiny of the legionariesin A.D. 14 and the beginningsof the deductio have al-
ready beentouchedupon in the previous chapter(p. 40). It is possiblethat the
uprising was responsiblefor the rapid building of the town; its wall was erected
eitherat the endof 14 or at the beginningof 15. It enclosesa rectanglemeasuring
522 by 435 m. The ditch, gatesand posternsof the town have beenexcavated.
Excavationscarried out by W. Schmid before the First World War revealed
someinsulae and the main featuresof the street-grid.This rectangularnetwork
incorporatesall the principles applied in the founding of veteransettlements.
The excavations,which have recently beenbegun,will probably supply much
informationconcerningthe earliestperiodswhich did not emergefrom previous
work. The extensivelyexcavatedcremationcemeteries,which have revealeda
wealth of material, extendedalong the road from the north gate, and have
provided evidencethat therewas a considerablesettlementfrom the beginning
of the first century A.D. The citizens settled in Emona on the basis of the
deductio were veteransof the legions stationedin Pannonia.It is mainly those
from XV Apollinaris that areattested,thoughthereis an inscriptionto a veteran
from VIII Augusta. Most of the personsattestedby the earliest epigraphic
evidence,were, however, civilians; from this it may perhapsbe assumedthat
74
N
EMONA
o1
100
1
200
1
300
1
metres
SAVARIA
Figure I} Town plans I
75
civilians were also involved in the deductio. The civilians concernedwere mainly
north Italians, a few from other parts of Italy, and somesouthernGauls. There
were also many civilians from Aquileia; this is in itself not surprisingin view of
the importantpart this neighbouringtown playedin the commerciallife of the
Savevalley and the amberroad. Citizens involved in the deductio were allocated
plots of land in the upper Savevalley north-westof Emona,whereasthe nearby
marshland,south-eastof the town, was left to the native inhabitants.Evidence
of the former is providedby numerousinscriptionsfound roundthe present-day
village of Ig. By reasonof its position Emona was destined to become an
entrep6tand the most importantcentrein Pannoniaof the commercialnetwork
basedon Aquileia. This may well explainwhy the group of citizensfrom Emona
so soonappearedin Savaria.The traders'settlementand entrep6tat Nauportus,
establishedunder the late republic, soon lost its importanceand becamea vicus
within the territory of Emona,despitethe fact that in A.D. 14 it was a settlement
municipiiinstar.
Savaria (Szombathely)lies in a fertile plain on the western edge of the
province. Both soil and position madeit much more suitablefor a deductio than
was Emona,which hadthe soleadvantageover Savariathat it was a communica-
tion centre. In the choice of site for founding a colony the main criterion was
probably the needto establisha strong-pointon the long stretchof the amber
roadbetweenthe legionaryfortressesat Poetovioand Carnuntum.It is probably
not fortuitous that the Tiberian colony of Emonawas roughly midway between
Aquileia and Poetovio, and the Claudian colonia roughly midway betweenthe
latter and Carnuntum.The deductio of veteransunder Claudius at Savariathus
fits into the plan which originated at the time of the initial occupationof the
country.
Beforethe time of Claudiusthe territory of the colonia ClaudiaSavaria,along
with many other parts of Pannonia,belongedto Noricum. Velleius emphasizes
that in A.D. 6 Carnuntumwas 'a place within the regnum Noricum', while Pliny
introduceshis descriptionof Pannoniawith a brief statementabout the deserta
Boiorum, adding'but now in the desertathereare the settlementsof the oppidum
ScarbantiaIulia and of the Claudian Colonia Savaria'.83 The Amber Road on
the western edge of Pannoniaand on the easternedge of Noricum was thus
probably treatedas a special area (Noricis iunguntur deserta Boiorum); its native
inhabitantssharedin the Julio-Claudianrights of citizenship grantedto Nori-
cum, indeedeventhe deductio of the colonia Savariacould easily havefitted into
the Claudianmunicipalizationprogrammedevisedfor Noricum. HenceMomm-
senstill held the view that it was only after Claudiusthat Savariawas attachedto
Pannonia.It was in fact only gradually during the first half of the first century
76
that placesalong the Amber Roadbecamepart of Pannonia:the exceptionwas
Celeia(Celje = 2illi) which remainedin Noricum. Evenin later timesthe stretch
of the Amber RoadbetweenthereandPoetovioran throughNoricum.
Veteransof legio XV Apollinaris were involved in the deductio of Savaria,a
fact underlined on tomb-inscriptions.84 There is no evidenceof a similar in-
volvement of the other Pannonianlegion, XIII Gemina, which had recently
beenstationedat Poetovio.The veteranswereallottedplots ofland in the town's
fairly extensiveterritory, as inscribed tombstonesset up by Roman citizens
with Italian namesto the west, north and south-westof Savariatell US. 85 They
were probably mainly descendantsof veterans.The earliestveteransmust have
hada considerablenumberof slaves,since,frequently,their tombstonesmention
severalfreedmen.The veteranSextusUttiedius Celer had at least six freedmen,
which presupposes aminimum of twelveslaves.86 Someof theveterans,however,
did not remainin Savariabut returnedto Carnuntumwherethey settledcloseto
their old garrison. This can be deducedfrom the fact that on tombstonesfrom
the camp cemeteryalong the streetstarting from the Porta Principalis Sinistra
someveteransof legio XV Apollinaris are mentionedas being from Savariaor
belongingto the Tribus Claudia of colonia Savaria.87 It is not known whether
theseveteranshad sold their plots of land or whetherthey had beendischarged
with a missio nummaria.It is, however,not improbablethat they had left Savaria
becausewhere the Amber Roadcrossedthe Danubeat Carnuntumthere was a
better chanceof work. The veteransin Savariaboughttheir slavesin the slave
marketat Carnuntum;a freedwomanbelongingto the veteranUttiedius was,for
example,called Carnuntina.In later times, too, Savariahad close contactswith
Carnuntum.
Another componentof the citizen-bodyof Savaria,and one not to be under-
estimated,was a number of north Italian civilians. Among its leading families
there was a preponderanceof Aquileian traders,principally the Caesernii,Canii,
Caesii and Petronii; a number of other citizens belongedto the tribe Velina,
which was the tribus of Aquileia. The most famous Aquileian family was that of
the Barbii, who by the secondhalf of the first centurywere alreadydecurionsof
Savaria,88a fact which calls for specialemphasis,sincethis family otherwisedid
not take muchpart in the commerciallife of Pannonia.The oppositewas true of
the other important family from Aquileia, the Caesernii,89who, starting from
Savaria, spread throughout the northern part of the province. Epitaphs to
citizens of Savariahave also beenfound in Aquileia,90 and a freedwomanof the
Petroniiin Aquileia was calledSavarina.9I
Theseinterrelationshipswithin the urbanaristocracyare attestedfrom all the
towns on the Amber Road betweenAquileia and Carnuntum.There is always
77
somethingattractive about discoveringpersonaland family connectionsfrom
the study of family and individual names(gentilicia and cognomina).To this end
thereis abundantepigraphicmaterialfrom the first two centuries.For example,
a certain L. Atilius Saturninushad his son, a soldier in legio XlIII Gemina,
buried on the southernboundaryof the territory of Scarbantia.Now thereis a
freedmanwith the samenameburiedin Aquileia whosetombstonestatesthat he
was a citizen of Scarbantiaand had beenmurderedby latrones-probablyon a
businesstrip to Aquileia.92 Traders(negotiatores)are mentionedon a numberof
tombstonesin western Pannonia (PI. 5b); these were mainly freedmen of
Aquileian trading families and bore the latter's names.93
As for Savaria,a group from Emonaarrived there within the framework of
thesecontacts,possiblybeforethe town was founded. Shortlyafterwardsa man
from Emona who was not even a Roman citizen (Lucius Maximi filius) was
buriedin Savaria.9 4 Most of the evidence,however,belongsto a later date.Both
in Savariaand Scarbantiainscriptionsdid not becomecustomaryuntil after the
middle of the first century; and trading contactsare also first attestedby finds
in the secondhalf of the century.It was, however,underTiberius, Caligula and
Claudius that trade was developed,when the group of Roman citizens was
emergingat the oppidumof Scarbantiaand colonia Savariawas being founded.
As far as can be established,the town of Savariaconsistedof a large quadri-
lateral some 700 by 550-600 m with a rectangularstreet-system(Fig. 13). The
town-walls were probably built at a later date. It is likely that they encloseda
biggerareathan the quadrilateralcolony of the deductio,but so far definite traces
of them havenot beenestablished.Excavationsto date have also failed to pro-
duce evidenceof the earliestbuilding period underClaudius.It is possiblethat
the town acquiredits urban characteronly gradually. The capitolium, probably
a very imposing structure,was not built until about the middle of the second
century; on the southernperimeterof the town, in the first century,therewere
still primitive huts built of sun-driedtiles. The sewage-system,water supplyand
streetpaving probably date back to the founding of the town by Claudius.The
streetswere pavedwith very large basalt slabs, while a subterraneanaqueduct
suppliedwater from a considerabledistance(c. 50 km); in order to ensurethe
necessarygradientthis conduit was adaptedto the terrain, and so to begin with
it ran from north to southand turned eastonly when it reachedthe altitude of
Savaria.9 5
At the latest when the veteransettlementwas founded under Claudius, and
possibly even earlier, the land of westernPannoniawas surveyedby centuria-
tion. Until recentlytracesof the field boundariesfixed in accordancewith all the
rules of the Romangromaticicould be detectedon modernmaps.96 The centuriatio
78
of the ager Savariensiswas uniform with this system, having a deviation of
22-23 degreesfrom north to west. The same orientation was applied to the
streetsof the town. The land thus centuriatedwas not all allotted to veterans.
The native inhabitants,particularly in the northern and westernparts of the
territory, were left undisturbedon their old properties;in many placesin these
areasthere are hundredsof tumuli, doubtlessnative burial places.This is con-
firmed by inscriptions to peregrini with Celtic namesand also by inscriptions
found in the tumuli.97 IndigenousCeltic namespersistedin theseparts into the
secondcentury; instancesof marriagesbetweennatives and citizens from else-
where are also attested.98Thus natives living in the open country were not
simply expropriatedin order to makeway for the settlementof veterans.Some
of them took part in the life of the town and were actually grantedcitizenship
underClaudius;they were not completelyrelegatedto the background.99 In the
pre-Flavianperiod a tombstone(PI. 5b) was set up to a certain Atta Bataionis
filius who was describedas a negotiator.lOO Attention has already been called to
the relatively rich storeof objectsfound in tumuli in west Pannonia;bronzeand
glassvesselsandsigillata potteryarenot particularrarities thereandindicatethat
therewas a tendencyamongthe native upperclassto adaptits way of life to the
materialculture of the foreign settlers.This processobviously took a fairly long
time. Evenin Savariaonly the frameworkof its urbancharacterwas established
under Claudius,and the useof importedRomangoodswas but the first step in
the transformationof the native way of life. Nevertheless,the cemeteriesso far
excavatedin Savariaindicate that the local production of objects of a Roman
type was sufficiently considerablein the secondhalf of the first centuryto oust
their own native productS.lOI This is in striking contrastto what is found in the
northern and easternparts of Pannonia,where, in the same period, native
industrylargely dominatedthe marketand was also able to satisfy the demands
of the auxiliaries.
79
Chapter 4
The Danube frontier from Vespasian
to Marcus Aurelius
Experienceson the Danubefrontier in the Year of the Four Emperors,and not

leastthe defeatof FonteiusAgrippa, promptedthe new Flavian governmentto
revise the Augustanscheme.This does not mean that troops stationedin the
interior of the province were immediatelytransferredto the frontiers; on the
contrary,this processoccupiedthe periodfrom Vespasianto Trajan. Neverthe-
less, the occupationof the frontier areasby strongerforces and the building of
garrison-postswere regardedas urgenttasks.It is reportedthat the governorof
Moesia, Rubrius Gallus, orderedconstructionwork to be undertakenon the
Danube in the earliest years of Vespasian'sreign: I C. CalpetanusRantius
Quirinalis, governorof Pannonia,had similar work carriedout in that province.
He is mentionedon four building-inscriptions,with almostidenticaltexts. Three
of these,which werefound at Carnuntumandare datedto the year 73,2 refer to a
building of legio XV Apollinaris, whereasthe one from Aquincum3 was set up
by an ala, possibly ala HispanorumI; it is probably of the same date as the
Carnuntum stones. These inscriptions attest military buildings, both in the
fortress at Carnuntumand in the auxiliary fort at Aquincum; thereis, however,
no meansof establishingwhetherthesestationswere being rebuilt in stone.The
inscriptionfrom Aquincumis particularlybig (roughly 3 m in length); hencethe
building which carried it must have been of someimportancewithin the fort.
The older readingof this inscriptionled to a variety of theories:for examplethat
a stonefort hadbeenbuilt at Aquincumas early as A.D. 17-20andthat, therefore,
the legionary fortress of Carnuntummust likewise have beenof stone.4 Since,
however, it has recently been establishedthat it is not [T]i(berius) Caesar but
T(itus) Caesar who is mentionedon the Aquincum inscription, the date of the
80
The Danubefrontier from Vespasianto Marcus Aurelius
stone fort is again an open question. The inscriptions of the governorshipof
C. CalpetanusRantius Quirinalis, however,do not prove the erectionof stone
forts then at Carnuntum and Aquincum, as a stone building-inscription is
known from Carnuntumdatedto the years 53-4.5 However this may be, these
inscriptions of Calpetanusfit into the framework of the new Flavian frontier
policy. This canbe illustratedwith somefurther details.
First of all we haveproof of the existenceof the fort at Taliata (Donji Milano-
vac) in the Djerdapin the early yearsof Vespasian. Recentlya military diploma
issuedon 27 April 75 6 to a cohort soldier from Antioch was discoveredthere.
He had presumablysettlednear his old garrison.Apart from this, we have no
definite evidencethat auxiliary forts were establishedon the Danube.Flavian
military diplomasprove that it was under Domitian that the auxiliary forces in
Pannoniaand Moesia were first considerablyreinforced. Despite this there is
evidencethat fresh auxiliary troops were arriving underVespasian:for instance,
cohorsI Noricorum equitata,which took over the fort at Arrabona,7cohorsI
Thracum Syriaca,which was now stationedat Timacum minus (Ravnain the
Timok valley),8 and perhapsalso cohors I Raetorum,which may have been
transferredto the Danubefrom Syria along with cohorsI Thracum Syriacain
the expeditionof Licinius Mucianus.9 The stationingof cohorts of infantry on
the Danubefrontier perhapsindicatesa changein tactics, as previously it had
always beenheld by alae. In the absenceof further excavationit is impossibleto
say which Danubeforts were constructedunderVespasian.Generallyspeaking,
however,it is very probablethat someat least were built for the newly arrived
cohorts. It is possiblethat there was some connectionbetweenthem and the
reorganizationof the Danubefleet. There is, of course,evidencethat a classis
was alreadyin existencein the first half of the first century, yet both the Pan-
nonian and the Moesianfleets were termedF favia.
The new plan for guardingthe Danubefrontier took the form not so much of
a suddenincreasein the forces in the provincesas of an alterationin their de-
ployment. Vespasiantransferreda considerablenumber of auxilia from Pan-
nonia to the Rhine, amongstthem ala Asturum II, ala Scubulorumand cohors
XXXII V oluntariorumc. R.; there is no evidencethat such large numbersof
troops arrived under him as under Domitian.The first indications of the new
frontier policy arethe increasingrole of the cohortsandthe strengtheningof the
occupationforces in the frontier zone. In Moesia the catastrophicevents of
69-70 made radical intervention necessary.The only legion (VII Claudia) in
the areaof what was later Moesia Superior proved inadequatein face of the
Dacian danger. Accordingly, Vespasian transferredlegio V Alaudae there.
Whereit was basedis not known; that it was somewherein Moesia in the early
8r
Flavianperiodis attestedby an inscription on a tombstoneset up to a veteranat
Scupi.10
The military precautionstakenby Vespasianensuredpeaceon the Danubefor
a decadeand a half. Thereis no evidencethat hostilities broke out either under
Titus or in the early yearsof Domitian. It was, however,during this period that
the reunified Dacianstatecameinto being. The Dacians,who had risenin force
in 69-70 against Rome-an indication that this people was recovering its
strength-hadbeen united, or rather reunited, into a single state by King
Diurpaneus,the predecessor of Decebalus.The centreof his state,the royal seat
at Sarmizegethusa, lay dangerouslynearthe Danubefrontier ofMoesia.
It would seemthat the outbreakof the Dacian wartook the Romansby sur-
prise. The first attack on Moesiaat the end of 85 or the beginningof 86 was so
unexpectedthat the governor,Oppius Sabinus,togetherwith a legion, probably
V Alaudae, was defeated.II Domitian, accompaniedby Cornelius Fuscus,the
praetorianprefect,went in personto Moesia. In succeedingyearsall the neces-
sary conclusions were drawn: the provincial armies on the Danube were
increasedto several times their original strength; most of the legions were
transferredto the bank of the river; commandof the long Danubefrontier in
Moesia was subdivided. These measuresinitiated a very troubled period of
reformswhich lasteduntil Hadrian.Researchhasby no meansreachedthe point
where it is possibleto outline the history of troop movementsin thesethree
decades.While thereis no dearthof hypotheses,lack of cogentevidenceand of
reliable cluesis equallyimpressive.
Domitian's initial measuresaffected the threatenedprovince of Moesia. To
replace the annihilated legio V Alaudae, IlIl Flavia was transferred from
Dalmatia, and from the west two more legions, I and II Adiutrix, were moved
into the province. Thesegreatly strengthenedforces were divided betweentwo
commands:from 86 the Danubefrontier as far as a point eastof Ratiaria was
assignedto the newly createdprovince of Moesia Superior, while the lower
Danubedown to the Black Seabelongedto MoesiaInferior. From the start the
governors of the two new provinces were consulars.The first governor of
Moesia Superior,L. FunisulanusVettonianus,had beengovernorof Pannonia
and probablyarrived in the threatenedprovinceimmediatelyafter the defeatof
OppiusSabinus.In orderto unite thewhole of the Daciansectionof the Danube
frontier under one commandthe region of Sirmium was attachedto Moesia
Superiorin 86.12 It is usuallyassumedthat, of the newly arrived legions,I and II
Adiutrix were stationedin the neighbourhoodof Sirmium and IIlI Flavia was
in the samefortressas VII Claudiaat Viminacium. This was probablythe head-
quarters.Ratiaria and Singidunumare also possibilities.Moreover, we cannot
82
exclude the possibility that not all the legions were stationedon the Danube
frontier of Moesia Superior; one might well have been assignedto Moesia
Inferior. From a geographicalpoint of view it is difficult to understandwhy
Ratiaria,separatedas it is from Moesia Superiorby the StaraPlaninamountain
chain, was not attachedto the lower province.If, however,we can assumethat
one legion was stationedat Ratiaria, and the two legions already in Moesia
Inferior (I Italica and V Macedonica)were augmentedby a third, we can then
provide an explanationwhy Ratiaria was assignedto Moesia Superior. For in
this way eachof the Moesianprovinceswas allotted threelegions. If this is so,
then there could have beenonly one legion at Viminacium. In any case,double
fortresseswere at this time increasinglycoming to be regardedas unsatisfactory
becauseof the dangerthey offeredto internalsecurity.
In the same year, 86, Domitian handed over commandto the praetorian
prefect, Cornelius Fuscus, and returned to Rome. The province was being
clearedof Dacians, but in his subsequentexpedition Fuscuswas disastrously
defeated,I3losingboth his life and his army. Domitian onceagainhurried to the
seatof war. The commandwas entrustedto Tettius Iulianus, who finally won
the battle of Tapaein 88. Decebalus,the Dacianking, suedfor peace.I4
At first sight it seemssurprisingthat Domitian did not follow up his victory,
or at least exercisestricter control over an enemyresponsiblefor the defeatof
one legion and the deathof three Roman generals(FonteiusAgrippa, Oppius
Sabinus and Cornelius Fuscus). He remained content, however, with one
Dacianvictory and immediatelymovedagainstanotherenemy,the Germanson
the Danube.Thereis much to supportthe view that in the early 80S he had al-
ready beenmobilizing againstthem; at that time he was alreadyconcentrating
troops on the GermanDanubefront, IS and it is not out of the questionthat the
sudden appearanceof three legions in Moesia in 86 was made possible by
previouspreparationsfor an assaulton the Germans.The Dacian attack came
as such a surpriseto the Romanssimply becausethe governmentwas concen-
trating all its attention on the Germans.This had probably not escapedthe
notice of Decebalus.
The causesof the threat of war on the Danube frontier of Pannoniaand
Noricum are not known. There is no further mention of Sido and Italicus, the
kings of the Quadi who had proved loyal vassalsof Rome during the Year of
the Four Emperors,beyondthe fact that they took part in battles in Italy on
Vespasian'sside and were involved in the plundering of Cremona; nor do
sourcesmention any changeof dynasty. According to Tacitus both Sido and
Italicus were hated by their people, and he adds that both remainedloyal to
Rome.I6 By the 80S both were probably already dead, as they had, of course,
8}
The Danubefrontier from Vespasianto Maretls Aurelius
beenreigning since 5o. By the time of Domitian they must have beenadvanced
in years.According to CassiusDio the immediatecauseof war was the refusal
of the Germansto supply Domitian with military aid for his DacianwarP Such
a demandwas by no meansunusual:Sido andItalicus had beenrequiredto take
up arms on the side of the Flavians in the civil war of 69. Thus the Roman-
Germanallianceinvolved mutual aid in the eventof war. Whenthe Marcomanni
and Quadi refusedto supply troops, Domitian attackedthem and actually had
their peace-emissaries put to death. Theseeventsmust have taken place in the
monthsfollowing the victory at Tapae,at a time when Domitian's handswere
somewhatfreer. In order to engagethe Germansthe Romantroops had to take
the shortest route from the Dacian theatre of hostilities; one inscriptionI 8
indicates that some of them had to march through Dacia, i.e. northwards
throughthe Hungarianplain. Episodicaccountsavailablepermit no satisfactory
chronological reconstruction,nor do they provide any indication of results.
Domitian soon left Pannoniafor the Rhine, but had to return immediately(in
89) as Romantroops had suffereda defeat. It was at this point that peacewas
formally concluded with the Dacians; Decebalus,who did not wish to be
presentin person, sent his brother Diegis to the negotiationsas his plenary
representative.He receivedthe diadem from Domitian and appointmentas a
client-king. Consideringthe tensesituation in Pannonia,the peaceterms were
very mild.I 9
In Rome in the sameyear Domitian celebrateda triumph over the Dacians
andthe Chatti, but not over the Germanson the Danube.This suggeststhat the
war againstthem in Pannoniawas still in progress.There is, however, no in-
formation concerningeventsin that provinceimmediatelyafter 89. It is probable
that Domitian succeededin winning over northern neighboursof the Marco-
manni and Quadi to the Roman cause,thus enabling him to attack the Suebi
later from the north. Not long afterwardsthe Lugii appealedfor Romanhelp
againstthe Suebi; in reply Domitian sent them formal help in the shapeof a
hundredcavalrymen.The result was that the Suebi then involved the Iazyges
on their side, and hostilities, which-probablyby diplomatic means-Domitian
had managedto postponeafter 89, now broke out as the Suebian-Sarmatian
war.20
Thereis very little information aboutthis war. In the work of later historians
there is, however, frequent mention of a catastrophicdefeat, which attracted
attentionnot only becausea whole legion was destroyed,but also becausethis
defeatcould be laid at the door of Domitian.21 The legion, possibly XXI Rapax
which had recentlybeentransferredfrom the west, was defeatedin Pannoniaby
the Sarmatiansand this causedDomitian to return once again to the province,
84
The Danubefrontier from Vespasianto Maretts Aurelius
where he remainedfor eight months in either 92 or 93.22 On this occasion
he did not return to Rome unsuccessful,for he celebratedan ovatio over the
Sarmatians.
The modestovation, insteadof a triumph, suggeststhat Domitian still had
further plans in view in Pannonia.It would seemthat the Germanswere not
finally defeated.The fact that the war had no definite name in contemporary
terminology implies that it was not regardedas finished.23 Indeed there are
reports of a Suebianwar undertakenby Nerva. The end of the war coincided
with the adoptionof Trajan.24
The Suebian-Sarmatian wars of Domitian were of decisiveimportancefor the
military situationin Pannonia.As has beenseen,the auxiliary forces had already
been reinforced before the outbreak of the war, a fact which is attested by
military diplomasfrom the years 80, 84 and 85. Oncethe war had startedit was
necessaryto bring in a legion. This was XXI Rapax,which was soondestroyed
and was immediatelyreplacedby XlIII Gemina.Thereareindicationsthat these
two legions were basedin the south-eastof the province, near Mursa.25 As a
result of the transferenceof XXI Rapax, or rather of XlIII Gemma,to Pan-
nonia, this becameonceagaina three-legionprovince. A little later the army of
occupationwas reinforced by a further legion, and so the position remained
until late antiquity. The Dacian and Suebian-Sarmatian wars under Domitian
resulted in Pannoniabecoming once and for all one of the most important
military provincesof the empire.
At the sametime the difficulties of tracing the distribution of the Danube
legions under Domitian and Trajan must not be glossedover. It is known, of
course,that the occupationarmy in PannoniaandMoesiawas almostdoubledat
the outbreakof the wars, and that as a corollary the Rhine army, from which the
troops transferredto the Danubehad in the main beendrawn, lost its primacy.
It is also possibleto list the legions and a considerablenumberof newly arrived
auxiliary units by name; what is, however, completelyuncertainis where they
were based.Legionaryfortresses,including doublefortresses,in addition to the
old attestedonesat Carnuntumand Poetovio,may well haveexistedat the time
of Domitian and Trajan at Brigetio, Aquincum and Mursella in Pannonia,with
anotherin the neighbourhoodof Sirmium; and in Moesia Superior,as well as
the known Viminacium fortress, at Singidunum,Ratiaria and possibly also at
Naissus.The only late Flavian to Trajanic fortressesto be partly revealedby
excavationarethoseat Vindobona,Brigetio andAquincum. Butevenwith these
it has not beenpossibleto determinewhen they were built. This is all the more
regrettable,sinceit is clearfrom the start that the legionsmarshalledagainstthe
Dacians also took part in the Suebian-Sarmatianwar without exposing the
85
Dacianfront. At this point a factor which is often ignoredmust be underlined:
as we haveseen,after the outbreakof the Dacianwar the region of Sirmium was
attachedto Moesiaand only after the defeatof the Daciansin Trajan'sreign was
it returnedto Pannonia.It is also definitely known that the westernboundary
of Decebalus'Dacian empire was the river Tisza; the greaterpart of the Sir-
mium region was, therefore,oppositeDacia. It is, then, very probablethat part
of the Danubearmy was stationed near Sirmium; evidenceof this is providedby
the great number of late Flavian and Trajanic inscriptions discovered in
Sirmium which refer to Danubelegions (IlII Flavia, I and II Adiutrix).26 Some-
where in the Sirmium region, which was now playing its accustomedleading
role, a fortressmust certainlyhavebeenplaced,but its precisesite is unknown.
The coloniaSirmium, foundedunderVespasian,is the leastlikely place.
It is fairly certain that legio IlIl Flavia was transferredto Aquincum im-
mediatelyafter the endof the Dacianwar, that is in 89 or shortly afterwards,and
beganbuilding the fortress there. Its camp prefect, L. Naevius Campanus,had
altars erected,and it is probablethat a big Domitianic building-inscription27
was cut in the same stonemason'syard. The fortress at Aquincum was built
north of the late La Teneoppidumon Mount Gellert at the point wherea north-
west to south-eastvalley reachesthe river-bank(Fig. 14). This valley provideda
good connectionwith the west-eaststretchof the Danubefurther upstream;a
diagonal road by-passingthe river-bend north of Aquincum led to the porta
decumanaof the fortress. The chronologicalrelationshiphere betweenlegiones
IIIl Flavia and II Adiutrix is still uncertain. There arecertainly traces of the
latter at Aquincumdatingto the late Flavianperiod.28 Perhapsit was represented
only by a vexillation or perhapsit was transferredtherefor a time beforethe out-
breakof the Germanwar and was only later movedto the Sirmium region. The
problem is madeeven more difficult by the fact that soldiers belonging to II
Adiutrix wereburied southof the fortressand at a considerabledistancefrom it,
at a spot wheretombstonesof auxiliaries have also beenfound. On the basisof
thesetombstonesand the large numberof samiansherdsof the end of the first
centuryfound in the area,it is assumed,perhapsrightly, that therewas another
Flavian fort there (Fig. 14) which, becauseit was sited on a narrow partof the
river-bank,couldnot havebeenlarge enoughto accommodatea whole legion.29
In view of the difficulties with which Domitian had to contendon the Danube
frontier of Pannoniabetween89 and 93 (particularly in connectionwith the
formerly peaceableIazyges)it is by no meansimprobablethat for a time Aquin-
cum had manytroops.Within the Carpathianbasinit did, of course,provide the
bestcrossing-pointof the Danube;and it shouldnot be forgotten that up to the
destructionof the independentDacianstateby Trajan the main settlementarea
86
Amphitheatre @)
Amphitheatre @) Amphitheatre
Amphitheatre @)
Amphitheatre @)
eatre
Amphitheatre @) Amphitheatre @)
Amphitheatre @) Amphitheatre @)
Amphitheatre @)
Amphithe
Amphitheatre
Amphith
o, o 2, 3
I
Late
tre )
ea @
)
@
ith tre
Amphith Late Ron1an Fort

ph hea
EskU'-ter
Am phit
Am
Amphitheatre @)
Figure I4 The environsof Aquincum

The Danubefrontier from Vespasianto lvIarcus Aurelius
of the Iazyges was in the north-westernpart of the Hungarianplain, exactly
oppositeAquincum.3o
The legionary fortressesat Vindobona and Brigetio were built at the latest
aroundthe turn of the century.A building-inscriptionfrom Vindobonais dated
no later than 103,3I but it is probablethat legio XIII Gemina,transferredfrom
Poetovio,was alreadystationedthere before this. It left behinda large number
of stampedtiles and building blocks32 which can be preciselydated,particularly
as this legion was attachedto Trajan'sexpeditionaryarmy in the Dacianwar and
remainedin Dacia thereafter.Legio XIII Geminaalso took part in the build-
ing of the fortress at Brigetio. The tiles stampedVEXIL III refer to vexilla-
tions of three legions, which, as other stampsfound in Brigetio tell us, were
XIII Gemina, XlIII Gemina and XV Apollinaris. After the departureof the
first the work was continuedby the other two legions; thus, a very early date
is unlikely for the start of construction.33
As a result of the transferenceof XIII Geminafrom Poetovioto the Danube
and the abandonmentof its former fortress, thearmy in Pannoniawas concen-
tratedon the river, and the interior of the provinceremaineddemilitarizedfor a
long time. As far as presentknowledgegoes,all the later auxiliary forts on the
Danubewere built between89 and 102, with the result that there was a dense
chain of them along the river. From the interior of the province, on the other
hand, we have no evidencefor auxiliary or other stationsduring the secondor
third centuries.Later Romanhistoriansplace the establishmentof the Danube
limes in Pannoniaprecisely in this period and connect it with the Suebian-
Sarmatianwar.34
At present,Flavian auxiliary forts are certainly attestedonly at Vetus Salina
(Adony), whereL. Barkoczi discoveredfour earth-and-timberforts with round
towers.35 He is also responsiblefor the very reasonabletheory that Flavian
auxiliary forts should be sought where so-calledPo valley sigillata pottery has
beenfound. Furtherclearly interpretableevidencemakesit possibleto draw up a
list of Flavian auxiliary forts, though it cannot, of course,be consideredcom-
plete. Auxiliary forts either built under the Flaviansor alreadyin existencemay
be assumedat the following places:Cannabiaca(Klosterneuburg),Ad Flexum
(Magyarovar), Arrabona, Brigetio, Albertfalva, Vetus Salina, Lussonium
(DunakomlOd), Ad Militare (Kiskoszeg-Batina), Teutoburgium (Dalj) and
Malata (Banostor).36
It is very probablethat the limes of Moesia Superiorwas also consolidatedin
the late Flavianand Trajanic period. The importantrescueoperationswhich are
being carried out in the Djerdap will doubtlesssupply the evidence;at present
only interim reports are available, which indicate that the earliestlayers of the
88
forts there date from the end of the first century. A fragmentary Trajanic
building-inscriptionfrom the fort at Prahovo(Aquae?)37 must be datedto the
early years of this emperor,since it is unlikely that after the conquestof Dacia
there was any needfor new construction-workon the stretch of the Danube
which ceasedto be an imperial frontier.
The extensivebuilding activity on the Danubeline in Pannoniais not attribut-
able to any danger from the Dacians. The decision to double the army of
Pannoniafrom two to four legions, and to concentratetroops on the Danube,
probably arosefrom the recognition that Augustus'old plan of defendingthe
frontier by diplomaticmeanswasno longerviable. The newFlaviangovernment
reachedthis conclusion at the outset and took steps which would not have
suited the Augustan scheme.A basically different frontier policy vis-a-vis the
DanubeGermansand Sarmatiansprobablyformed part of Domitian'splans,for
even before the outbreak of the Suebian-Sarmatian war he had concentrated
troops in Pannonia. The killing of the Suebian peace-emissaries revealed a
radical breakwith the old methods,and the bitter experiencesof the war which
ensuedwent further to convince the emperor that security for the imperial
frontiers was only possiblefrom a position of strength,that is to say from a
constantdemonstrationof strength.
We do not know what kind of peace-treatywas concludedwith the Suebiand
the SarmatiansunderNerva. In any caseTrajan did not revertto the old scheme
but had constructionwork on the limes continued.The legionary fortressesat
Vindobonaand Brigetio were apparentlybuilt in the very earliestyears of his
reign.38 He was, nevertheless,concernednot to aggravatethe tensesituationin
the Carpathianbasin.In the Germania,written inthefirst yearof the reign,Tacitus
relatesthat theMarcomanniandthe Quadi had kings by the graceof Rome. The
samesourcestatesthat the old dynastyof Tudrushadbeenreplacedby 'foreign'
kings. Thesekings were installednot later than 98, when Trajan, who had just
beenappointed emperor, madeadetourto theDanubebeforereturningto Rome.3 9
There is a good deal in favour of the view that the much debatedRoman
buildingsin SlovakiaandMoravia (Fig. 15) arelikewise connectedwith Trajan's
treatieswith the Germans.As for their date, we can at least be certain that the
beginningof the secondcentury, when legio XV Apollinaris was still stationed
in Pannonia,is the latest possiblecontext, for this legion provided tiles.40 The
tiles, coming as they do from military units, mainly legions, have prompted
scholarsto regard thesebuildings as Romanforts, or at least as the housesof
commandersof Roman troops. It is known that, in the last years of Marcus,
Romanunits werestationedin Marcomannia,and from the time of Valentinian
therearealsoreportsof Romanforts in the territory of the Quadi.But apartfrom
89
Bath
MUSOV OBERLEIS
50 metres
MILANOVCE
STUPAVA
Musov
Oberleis
OBERLEIS
ROMAN BUILDINGS
IN THE OBERLEIS
OBERLEIS
BUILDINGS
BARBARl~VM
OBERLEIS
Figure If Romanbuildings in the Barbaricum

the tiles from military tile-works, thesebuildings have nothing military about
them. They were simple dwelling-houses,furnishedin the Romanmannerwith
a certaindegreeof comfort, and stoodin a courtyardsurroundedby a wall. The
walls were some50-60 cm thick, but of a rampartor ditch thereis no trace.Had
they beeninside the province, thesebuildings could easily have beentakenfor
small villas; it is impossibleto regard them as the fortified housesof military
commanders,still less as strong-pointsor forts for troops. Someyears ago E.
Swobodacame up with the idea that they were built by Roman troops for
barbarianchieftains; there is nothing odd about this as, after all, accordingto
Tacitus, the power of these chieftains rested on the auctoritas Romana,which
supported them financially-raro armis saepiuspecunia. Why should not the
Romansalso havebuilt residencesfor theseclient-princes?41
At the beginningof Trajan'sreign, if not before,a new treaty was concluded
with the Sarmatians(Iazyges),for during Trajan'sDacianwar they providedthe
Romanswith troops, and immediatelyafter it was over invoked certainagree-
ments.42Trajan's attention was, however, primarily directed to the Dacians.
Even after the peaceof 89 tensionon the Moesianfront was not fully relaxed.
Dio Chrysostom,who hurried therefrom his partly self-imposedexile in Dacia
after hearing the news of Domitian's death, found himself in the middle of
preparationsfor war. The descriptionin his Olympic Discourse43 refers to a big
camp in Moesia, probably Viminacium: 'There one could see everywhere
swords, everywherecorselets,everywherespears,and the whole place was
crowdedwith horses,with arms, and with armedmen. Quite alone I appeared
in the midst of this mighty host, perfectly undisturbedand a most peaceful
observer of war ... desiring to see strong men contendingfor empire and
power and their opponentsfor freedomand their native land.' Around the year
100 large numbersof military units were concentrated in Moesia Superior.On a
diploma of 93 threealae and nine cohortsare mentioned,whereasa diploma of
100 refers to no fewer than twenty-one cohorts.44Preparationsfor war, of
course,involved troop movements.In the apparentconfusionof moving whole
legions here and there one principle emergesto which Trajan's supremecom-
mandadhered,andprobablynot without reason.Legionsnewly movedin from
the west were stationedin fortressesin Pannonia,whereasthose which had
alreadybeentherefor a fairly long time weresentinto battleagainstthe Dacians.
This in turn broughtabouta changein the distribution of legionsin Pannonia,
and to a lesser degreein Moesia Superior: in the latter province legio VII
Claudia, alreadyat Viminacium, remainedthere, and IIII Flavia moved to its
final baseat Singidunuma few yearsafter the end of the war. During Trajan's
Dacianwars Moesia Superiorwas, of course,the launchingareafor all military
91
operations,and it is probablethat the emperorestablishedhis headquartersat
Viminacium. Of the Pannonianlegions,IIII Flavia, XIII and XlIII Geminaand
XV Apollinaris joined Trajan'sexpeditionaryarmy, thoughit is morelikely that
the last two were representedonly by vexillations. Legiones I and II Adiutrix
were also involved in the Dacian war. Both X Gemina and XI Claudia were
moved from the Rhine as replacementsfor the legions sent to Dacia, but this
probably did not happen before the end of the first Dacian war. Legio XI
Claudiawas quarteredin the recentlybuilt fortressat Brigetio,45 and X Gemina
at Aquincum, where it left behind many traces of its building activities, and
where some of its soldiers are attestedon tombstones. 46 Evidence that this
legion was stationedat Aquincum over quite a long period is provided by the
fact that its veteranssettled near the fortress.47 After Dacia was conquered
legio XI Claudiawas transferredto Oescusin MoesiaInferior and XXX Ulpia
Victrix, a legion recently raised by Trajan, took over the camp at Brigetio.
Legiones XlIII Gemina and XV Apollinaris returned to Vindobona and
Carnuntumrespectively.
It is very possible that the division of Pannoniainto two provinces under
Trajan (Fig. 16) was directly connectedwith the distribution of the legions. In
102 the province was still undivided.48 In 107 Hadrian, the later emperor,is
attestedas praetorianlegateof PannoniaInferior; 49 the division mustthenhave
occurredbetween103 and 106. However, it is not certain that he was the first
governorof PannoniaInferior; it is not impossiblethat he was the first praetor-
ian governorbut followed eitheroneor two consulargovernorsof the province,
and that thereforethe province mighthave beendivided as early as 103. If so,
the immediatecauseof the division may havebeenthe presenceof a fifth legion
in Pannonia,the recently raised XXX Ulpia. The first governor of Lower
Pannoniawould accordingly have been a consular,becauseof the presenceof
two legions in his province. All this, however, is at presentpure speculation.
Nor is it necessaryto supposea purely military reason.After the division of the
two other big frontier provinces-Moesiain 86 and Germanyin 9o-Pannonia
remainedthe only provincewith a disproportionatelylong externalfrontier and
a disproportionatelylarge army. In addition, the Sirmian sectionof the frontier,
betweenthe mouths of the Drave and Save,which had beenaddedto Moesia
Superiorin 86, was returnedto Pannoniain 106 at the latest.50 The Pannonian
sectionof the frontier, 700 km long, faced two barbariantribes which not only
belongedto different linguistic groupswith fundamentallydifferent ways of life
but also requiredbasically different tactical approachesin both war and peace.
The experienceof a century of treaty-relationshipshad amply shownthis to be
so. The division of Pannoniameantthat the Upperprovincehadthe Germansas
92
The Danubefrontier from VespaSiaJ2to Marcus Aurelius
MARCOM
ANNI
VINDOBONAO Coloniae
Coloniae
CARNVNTVMO
Coloniae
BRIGETIO
AQVINCVM
en.
nia
»
no
Coloniae an
-» ;:0
fP
~
-»
no
:l>
isio
-i
-»
-»
nia
div
m
no
-» -
an
Coloniae
fP
:J>
-»
no
N
-<
isio
-» o
div
m
Coloniae en.
-»
-»
Coloniae
o Legions
o Coloniae
o, 100
t
200
I
km
Figure I6 The division of Pannoniaunder Trajan
neighbours,while the Sarmatians,always known as the Iazyges,borderedthe

Lower. Moreover,the new inter-provincialboundarywas drawn in sucha way
that the whole of the largely urbanizedwesternpart was assignedto Pannonia
Superior,and this province, in relation to its area,had a fairly short stretchof
frontier. PannoniaInferior, on the other hand, possessedonly a single urban
community at the time it was established,colonia Flavia Sirmium, and its
93
territory never achieveda width of more than 50 km; yet it carried a much
longer frontier zone. The legions were assignedin proportion to the relative
size of the provinces.PannoniaInferior was allotted one, whereasthe Upper
province, with three times the area, was given three. At the time of the
division the legions receivedrecruits from the provincesin which they were
stationed,and it is thereforenot unlikely that in fixing the provincial bound-
aries the question of recruitment was considered.Finally there was another
overriding principle which was consistently applied under Trajan in all the
Europeanfrontier provinces:the numberof coloniae correspondedto the num-
ber of legions in the province. In this connectionit shouldbe pointed out that
at the end of the first century Emona no longer belongedto Pannonia,but to
ItalyY PannoniaSuperiorhad three coloniae (Siscia, Poetovio and Savaria)and
three legions (XlIII Gemina, XV Apollinaris and XXX Ulpia), Inferior one
colonia (Sirmium) and one legion (X Gemina). Here, too, recruiting policy was
involved.
The rank of the governorsof the new Pannonianprovincesvaried according
to the numberof legions undertheir command;the governorof Superiorwith
his headquarters at Carnuntumwas of consularrank andwas given the province
immediatelyafter tenureof this office, while the governorof Inferior, basedat
Aquincum, was a praetorianand, generally speaking,becameconsul designate
while still legatus Pannoniae Injerioris. Hence it frequently happenedthat the
governor of Inferior soon becamegovernor of Superior, although it is not
possibleto lay down hard-and-fastrules aboutit. The governorsofInferior were
appointed fairly regularly from among the legionary legates in the Upper
province. That they were primarily viri militares occasionsno surprise.
The establishmentof the province of Dacia in 106 had an importancefor the
future of the Danubefrontier which must not be underestimated.The most
active and the least docile peoplewithin the Europeanbarbarianarea, and the
only barbarianrace in Europethat had managedfrom time to time to establish
a stablestate,was now finally eliminated.Its role was takenby Sarmatiantribes
of horsemen,the Iazygesand the Roxolani, who, until the time of Trajan, had
generally played only a secondaryrole in Roman-barbarianrelations on the
Danube.Down to Trajan it was Romanrelationswith the Dacianswhich were
of supremeimportance,and it was only in this contextthat the Sarmatiantribes
acquiredsignificanceas possibleallies or buffer states.Iazyganhorsemenfought
on Trajan'sside againstthe Dacians,and thoseof the Roxolani on the Dacian
side againstthe Romans.5 2 In the first century A.D. the Iazyges were wedged
betweenthe Romansand the Dacians in the northern part of the Hungarian
plain-it is not unlikely that the Romanshad establishedthem there as a buffer
94
state-buta sectionof them was still in the plain of Oltenia in the time of Tra-
jan.53 WhenDecebalusconqueredthis areathe Iazygesinevitably becamehostile.
They thereforefought on the Romanside, thoughfifteen yearsearlier they had
beenresponsiblefor the disasterof legio XXI Rapax. After the defeat of the
Daciansin 106 the Iazygesnaturally demandedthe return of Oltenia.54 But as
Trajanwas determinedto constituteDaciaa province,it was out of the question
to cedeOlteniato a barbarianpeople.The resultwas war on the Danubefrontier
ofPannonia,wagedsuccessfullyby Hadrianduring his governorship.55Details
of the peace-treatyare lacking, but it is very probablethat Romewas inclined to
accedeto some of the demandsof the Iazyges, and that the latter were easily
satisfied.The distribution of early Iazyganarchaeologicalmaterialindicatesthat
in the first century this tribe of horsemenhad occupiedonly the northern, or
more precisely the north-western,part of the Danube-Tiszaplain.56 Under
Decebalusthe Tisza was the westernboundaryof the Dacianstate.57 Oncethis
statehad disappearedin 106 a power vacuumarosein the areaeastof the Tisza.
Trajan did not incorporatethe plain betweenthe river and the Transylvanian
mountainsinto Dacia but left it to the Iazyges.Archaeologicalfinds from the
secondcenturyindicatethat during that periodthey conqueredthe whole of the
plain on both banksof the Tisza; vesselsfound in someof the so-calledIazygan
cemeteriesrevealfairly strongDacianinfluence.58 It is not improbablethat Rome
gavethe Iazygesthe Tisza plain in return for Oltenia. A westerly migration of
the Iazygesfrom Oltenia through the newly establishedprovince of Dacia is
also not unlikely; half a centurylater they were allowed an unimpededpassage
through Dacia.
Here we havethe main problem which faced the Romansafter the establish-
ment of a province on the left bank of the Danube.The Dacian statehad, of
course,beeneliminated,but the mountedSarmatiantribesmenin the Carpathian
basinand on the lower Danubewere not peoplepreparedto enterinto a system
of treatiessuchas hadbeenmadewith the Germans.The new province,wedged
betweenthe Iazyges and Roxolani, was to some extent a bastion which held
them in check on two fronts, but at the sametime it was an obstacleto the
westerlymigration of thesepeoplesand hencea causeof constantunreston the
limes. Nor was the solution to the Iazyganquestion,which was reachedin 107,
a long-termsolution, for ten yearslater the Romansfaceda frontal attackby the
Sarmatiansboth westandeastof Dacia.
The military occupationof Dacia requiredof the imperial army somelegions
and a large number of auxilia. The earliest army of occupationnaturally con-
sisted of parts of the expeditionary force. Those Pannonianand Moesian
legions which had taken part in the expedition soon returnedhome and only
95
Xln Geminaand nn Flavia59 are later attestedin Dacia. It was in 110 that the
occupation forces were first reduced; in that year a fairly large number of
auxiliary regimentswere sentbackto the provincesfrom which they hadcome.60
A further sectionof the expeditionary armywas orderedto the eastwhen Trajan
startedhis Parthianwar in 114. The auxiliary units of the new army of Dacia
were largely drawn from the expeditionaryarmy and from the standingarmy
of Moesia Superior.At the sametime the numberof auxiliary units in the latter
province was considerablyreduced,for the establishmentof a province on the
other side of the Danubehad a secondconsequencewhich must not be over-
looked, in that the river ceasedto be the frontier of the empire in the stretch
betweenViminacium and the mouth of the Alutus (Olt)-in other words along
the extremelyimportant sectionof the Djerdap. The Danubeforts, built earlier
mainly under the Flavians, were presumablyevacuatedfor the most part, and
that stretchof the river, which until 106 had beenstrongly held by troops, was
demilitarized. Thereafterit was only on the relatively short stretchof the river
between Singidunum and Viminacium that Moesia Superior was contiguous
with barbarianterritory; yet this sector of the frontier had two legions, nn
Flavia in Singidunumand vn Claudiain Viminacium, and someauxiliary units.
Down-streamfrom Viminacium the chain of forts was thinned out to such an
extentthat one canno longer speakof a military frontier. The strong-pointsstill
held on the borderbetweenUpperMoesiaand Dacia were probablythe follow-
ing: Pincum (Veliko Gradiste), Cuppae (Golubac), Novae (Cezava), Taliata
(Donji Milanovac) and Egeta(Brza Palanka)wheretile-stampsand inscriptions
relating to auxiliary units suggestthat there were garrisonsin the secondand
third centuries.61 This list of forts, mainly thoseof auxiliary units, will probably
be supplementedwhen excavationsin the Djerdap are completed,but in the
meantimea few facts can be established.First of all, the extremelysmall number
of stationsis striking. They are mainly to be found east of Viminacium in the
upper Djerdap (Gornja Klisura) and extending as far as the mouth of the
Porecka,near Taliata. Its valley led into the mining district of the Pek and
Mlava valleys and was the only placein the Djerdapfrom which the interior of
the province could be reachedby road. The Gornja Klisura could easily be
reachedfrom the west by barbariansand was the reasonwhy this stretchof the
tow-path had to be strictly guarded. The stretch below Taliata was not so
strongly held: Egetahad to have a unit (cohorsI Cretum),62as the road which
by-passedthe greatDanubebendof the Kasandefile andthe Iron Gatesmet the
Danuberoad at this point. Down-streamfrom Egetathereare no definite traces
of second-or third-centuryauxiliary forts. As for the auxiliary regimentswhich
in those centuries were attachedto the army in Moesia Superior, the most
96
The Danubefrontier from Vespasial1to Marcus Aurelius
striking featureis their small number;63 their strengthwas just over half that of
the two legions (some 6,500 auxiliaries against somewhatmore than 10,000
legionaries),extremelyweak comparedwith auxiliary forces in Pannonia,Dacia
andMoesiaInferior. Therewas only a single cohorsmilliaria, and the forts listed
abovewere correspondinglysmall, as a rule not more than two ha (roughly five
acres),thoughmostly not more than one ha.64 The demilitarizationof the limes
in Moesia Superiordid not involve the interior of the province. It was in fact
the only province on the right bank of the Danubewhere the hinterlandcon-
tinuedto be occupiedin the secondandthird centuries;a post-Trajanicauxiliary
fort is attestedat Timacum minus (Ravna) in the Timok valley on the road
betweenNaissusand Ratiaria, which was garrisonedat least from the reign of
Vespasian,notably by cohorsI Thracum Syriacain the period from Vespasian
to Marcus Aurelius.65 Auxiliary units were stationednot only on the Danube
frontier of Moesia Superior(wherethere were somesevenauxiliary forts in the
secondcentury)but alsoin the interior, asis shownby the fact that the provincial
army included a total of sometwelve auxiliary units. The discovery of further
forts in the interior would not be surprising. The military occupationof the
interior was necessaryfor internal reasons.While problems relating to the
defenceof the frontier were not forgotten, they were of secondaryimportance.
Of more account was the fact that Timacum minus was situatedat the point
where the road from the Danube to Naissus, andultimately to the Adriatic,
joined the Timok valley. At this point the shortestroutefrom the lower Danube
to that seacould be blocked.
This east-westconnecting route through the Balkan peninsula began to
acquire significance immediately after the establishmentof the province of
Dacia and becauseof it; the stretch of the Danubewhich had been evacuated
was also important as providing a link betweenDacia and the Mediterranean.
The new province was cut off from the south by the very high range of the
southern Carpathians,nor was it readily accessiblefrom the west. The best
routeslinking it with the Balkansran along its boundaries-through the Banat
from Viminacium and through the Olt valley. From the Danubea valley also
openedto the north nearDierna(OrSova),but eventhis shortroute hadno good
connectionson the southbank. Transdierna(Tekija) could only be reachedfrom
the west along the cliff road in the Kasandefile andfrom the eastonly along the
Danubevalley. The Tabula Peutingeriana, which fortunately includes the road
network in Dacia, showsthree roadsenteringfrom Moesiawith starting-points
at Viminacium, Taliata and Egeta, none of which could have been actual
crossing-points.From Viminacium the road followed the southernbank until it
finally crossed the river near Lederata (Banatska Palanka); from Taliata it
97
utilized the cliff road through the Kasan defile as far as Transdierna,the first
point at which the river could be crossed,while from Egetathe road followed
the left bank to crossat Pontes(Kladovo), oppositeDrobeta(Turnu Severin).
In short,all the crossing-pointscould be reachedonly by using stretchesof road
alongthe valley,andthis led to brisk traffic alongthe river-bank.
The most important crossing-pointwas, of course,Drobeta, where Trajan
hadhis famousbridgebuilt by Apollodorusof Damascus.The work was carried
out by units from the first occupationarmy of Dacia, which left behind large
quantitiesof stampedtiles.66 Tradition, hostile to Hadrian,hasit that he hadthe
bridge'swoodwork removed;67 without casting doubts on the statement,one
may say that it is highly improbablethat it was not replaced.Be that as it may,
Drobeta was the most important crossing-point,simply becausethe parts of
Dacialying southor north of the Carpathianscould equallywell be reachedfrom
it.
The establishmentof Dacia as a provincecontributedto the unification of the
interior of the Balkans,especiallyMoesia Superior,with the rest of the empire.
CommunicationbetweenDrobeta or Viminacium and the Mediterraneanwas
availablealong the old waterwaysof eitherthe Danubeor the Save,but a direct
route which terminatedat Brundisium ran along the north-eastto south-west
diagonalthroughthe Balkan peninsula;this acquiredimportanceonly after the
conquestof Dacia. It startedon the Adriatic at Lissus (Leshja, Alessio),and ran
north of the ScardusMons via the Kosovo Polje,68following the Timok valley
until it reachedthe Danubenear Ratiaria. From there it followed the river up-
streamas far as Drobeta.It was perhapsno accidentthat it was at Ratiariathat
Trajan establishedhis only colonia in Moesia Superior,for it was along this road
from Lissus to Ratiariathat the earliesturban centresin the new provincewere
founded,at Ulpianum, Naissus,Timacum minus and Ratiaria itself.
Trajan's reign was the first important turning-point in the history of the
DanubeareasinceAugustus.After the transitionalperiod underthe Flaviansit
was now that the deploymentof both legionary and auxiliary garrisons,the
direction of the main lines of communicationand many other things too were
settledfor a long time to come. After Trajan there were no changesof critical
importancewithin eitherthe military or the tradingspheres,althoughit was not
until Hadrian that troop movementsceased.Some changesdid occur in the
distribution of the legions in the first yearsof the latter's reign, but thesewere
the result of Trajan's Parthian war. Unfortunately we do not know when
Pannoniaacquiredthe four legions which until late antiquity formed the back-
boneof its army; but it is probablethat theselegionsarrivedin time to takepart
in the battleswhich werefought on the Sarmatianfrontier in 117.
98
The legions which returnedfrom the Parthianwar were I and II Adiutrix,
which took over from XXX Ulpia at Brigetio and X Gemina at Aquincum
respectively.Of the legionsstationedin PannoniaunderTrajan, XV Apollinaris
and XXX Ulpia left the Danubeareafor good; legio X Geminamovedinto the
fortress at Vindobona and XlIII Gemina into that at Carnuntum.Legio IIII
Flavia, which in the early yearsafter the conquestof Dacia was attachedto the
army of the new province,probablydid not becomepart of the army of Upper
Moesia until the beginning of Hadrian'sreign. The distribution of legions in
Pannoniaand Moesia Superior from the time of Hadrian to late antiquity,
familiar to geographersand historiansof imperial times,69was as follows:
Vindobona legio X Gemina
Carnuntum legio XlIII Gemina
Brigetio legio I Adiutrix
Aquincum legio II Adiutrix
Singidunum legio IIII Flavia
Viminacium legio VII Claudia
Thus threelegionsfaced the Suebiin the Carpathianand Czechregions,and
threefacedthe Sarmatiansof the Carpathianbasin. The Upper Moesianlegions
were stationed,so to speak,as reservesin the middle of a long frontier which
enclosedthe Sarmati(Iazyges) on three sides. The greatestpressurefrom the
Sarmatianswas felt on the long sector of the Pannonianfrontier from the
Danubebendto the mouth of the Save;it was chiefly legio II Adiutrix and the
many Lower Pannonianauxilia which fought in the battles of the secondand
third centuries,and it would seemthat lessusewas madeof IIII Flavia and VII
Claudia. It is quite possiblethat the land which was wedgedbetweenPannonia
andDacianorth of the UpperMoesiansectionof the frontier hadto someextent
becomea military zoneunderthe commandof the governorof Daciaand,there-
fore, that the Upper Moesianlegions were not in contactwith the Sarmatians.
Even more did legio IIII Flavia act as a permanentreservein PannoniaInferior.
On every occasionthat legio II Adiutrix joined an expeditionaryarmy, it was
replacedat Aquincum by IIII Flavia; this explains why the latter legion left a
numberof monumentsbehindin the capital of Lower Pannonia.70
The conclusionthat the UpperMoesianfrontier and army did not playa key
role in fighting the Sarmatiansemergesfrom the eventsin the years II7-19. As
alreadystated,Hadrian'ssolutionduring his governorshipin 107 to the problem
of the Iazygeswas short-lived.The latter, possibly reinforcedby thosewho had
migratedfrom Oltenia, conqueredthe plain eastof the Tisza; in the courseof
this conquestthey doubtlesscameinto conflict with the 'free Dacians'who had
99
The Danubefrontier from Vespasianto j'vlarcus Aurelius
remainedoutsidethe Dacian province. Information about thesebattlesis lack-
ing, thoughthe result of this extensionof their powerbecameclearin the yearof
Trajan'sdeath,when simultaneouslythe Roxolani attackedthe Lower Moesian
limes and the Iazygesthat of Lower Pannonia.The joint action undertakenby
thesetwo tribes points to a commoninterestwhich doubtlesslay in the fact that
their communicationswith each other were hinderedby having the province
of Dacia wedgedbetweenthem. After defeatingthe DaciansTrajan had entered
into an alliance with the Roxolani; in I 17 the reasonthey gavefor the war was
that Romehadreducedthe stipendiumpaid themby treaty.71Thereareno sources
which throw light on the demandsof the Iazyges;they were, however,probably
more considerablethan thoseof the Roxolani, sincethey continuedto fight after
their allies had come to terms. In the attack mountedby thesetribes Dacia's
governor,C. Julius QuadratusBassus,was killed and Hadrianhad to go to the
seat of war,72 where he invested his trusted agent, Q. Marcius Turbo, with
exceptionalpowers over Dacia and PannoniaInferior.73 The Roxolani soon
settled down, and it is possible that Hadrian replacedtheir king and had the
former ruler internedin Pola after granting him Roman citizenship.74Marcius
Turbo held the commandin PannoniaInferior andDaciauntil 119; this suggests
that the union of the two armieswas necessarynot only on accountof the joint
actionof the Iazygesand Roxolani but alsoin order to continuethe fight against
the Iazygesalone.In otherwords,by 117-19the Iazygeswerealreadyon Dacia's
westernboundary,and had conqueredthe whole plain betweenthe Danubeand
Transylvania.75
The war ended with the Iazyges sending an embassyto Rome, where, it
seems,a client-treatywas concluded.76 This time the solution must have been
satisfactory,as for the next half century nothing is heardof the Iazyges.It was
during this period of peacethat the road throughtheir country was finally built.
A direct connectionbetweenPannoniaand Dacia obviously becamenecessary
after the latter was conquered,and Trajan had probably attemptedto build a
road running from west to east acrossthe plain. Aurelius Victor credits him
with having had a roadfrom Gaul to the Pontusbuilt by 'wild peoples'.77How-
ever,the wars of conquestwagedby the Iazygesbetween107 and I 17 arehardly
likely to have encouragedits use. After II9 and before the outbreak of the
Marcomannicwars underMarcus Aurelius it was a busy road, particularly as it
was possiblefor a long distanceto go by boat along the river Maros (Marisia,
Marisos). The entrepotwas at the mouth of the Maros, at Szeged,wherein the
courseof demolishingthe medievalfort in the last centuryRomanstonesculp-
tures,tiles and an inscription were uncovered.78 The inscription was to a certain
Mercator (Fig. 17), overseerof the station (zJilicus), who was superintendedby
100
HICVL
RVM
ERCATC
VILIC
Figure I7 Inscription of Mercatorfrom Szeged
the praefectusvehiculorum.7 9 So far the courseof the road has not beeninvesti-
gated;since,however,Szegedis a fairly well-defined site its direction will have
beenwest-east,startingfrom the auxiliary fort of Lugio (Dunaszekcso),for the
importanceof which there is other evidence,and running along the Maros as
far as Dacia. Thesepeacefulconditionsprobablyled to the abandonmentof the
greaterpart of the Banat by the Romans.It has recently beenestablishedthat
legio IIlI Flaviawas basedafter 106 at Berzoviain the easternpart of the Banat.80
However,it was soontransferredto its later baseat Singidunum.Its withdrawal
was perhapsthe result of the evacuationof the greaterpart of the Banat,which
was given to the Iazyges.This may havebeenthe part of Daciawhich, according
to somesources,was evacuatedby Hadrianshortly after he cameto the throne,81
an event that was, perhaps,brought about by the embassyof the Iazyges to
Rome.
Scholarshave recently connectedsome layers of ash found in the auxiliary
forts of Campona,Vetus Salinaand Intercisawith the destructioncausedby the
IazygesbetweenI 17 and I 19.82 Coin-hoardsand evidenceof fire in the legionary
fortressat Aquincum havealso beendatedto this period83-whetherrightly so,
it is difficult to say. It is equally uncertain whether the Marcomanniand the
Quaditook part in the war of I 17-19. The only evidencefor suchan assumption
is one mention in the Historia Augusta84 and one undated reference to an
101
expeditio Suebicaet Sarmatica. The referencein the Historia Augustastatesonly
that Hadrianhadinstalleda Germanking, but this may not necessarilyhavehad
any connectionwith a war, and indeedmay equally well refer to west German
tribes. The expeditiomentionedin the cursusof a certainsenatorwould be better
datedto the time ofMarcus.85
Definite evidenceof a crisis on the Germanfrontier of Pannoniacomesonly
in Hadrian'slast years.If the periodsof peaceon the Upper PannonianDanube
frontier are compared,it emergesthat betweenevery two wars or diplomatic
crises there was a period of peacewhich lastedfor a generation:(i) from the
time when Vannius becameking in A.D. 19 to his overthrowin 50; (ii) between
the latter date and Domitian's Suebianwar in 88-93, or more correctly until
Nerva's victory and the establishmentof the new order by Trajan in 98; (iii)
between98 and the outbreakof a Suebianwar in 136 or just before, and (iv)
betweenthe conclusionof peacec. 140/4 and the outbreakof war in 167; each
of these periods lasted three to four decades,i.e. a generation.This regular
repetition of crisesin the client statesprobably had fairly profoundreasons.In
the first placetherewere the dynasticchangesamongthe Germans,which led to
fresh negotiationswith Rome and to new hopesof a more favourableoutcome
on the part of the Germans.In Roman-Sarmatianrelations there were also
periods of peace,the only difference being, it would appear,that they lasted
longer.From the immigrationof the Iazygesat the beginningof the first century
A.D. therewere no fewer than eight or nine decadesuntil the war underDomi-
tian, and from the peaceof II9 until 167 almostfive decades.Solid evidenceof
joint action by the Germansand Iazygesis availableonly for the time of Domi-
tian. The system of client-states,then, stood the test even after the Danube
frontier had beenoccupiedby troopsin the Flavian-Trajanicperiod. Every war
endedwith the appointmentof a Germanking by the grace of Rome, and he
was able to guaranteethe peaceas long as his life lasted.
The war which broke out in Hadrian'slast yearsalso endedin an investiture.
The exactdateof its outbreakis unknown.Coin-hoardsdiscoveredon the Upper
Pannonianlimesinclude terminal coins from the years130/1, 132/4and 137, and
layers of ashin Aquincum are also regardedas being connectedwith this war.86
As will be seenlater, these terminal coins give no indication of the date at
which the hoardswere buried. One thing is certain,however:L. Aelius Caesar,
who had beensingledout as Hadrian'ssuccessor,was entrustedin 136 with the
joint governorshipof the two Pannonianprovinces.Accordingto his biography
in the Historia Augustahe had had a few military successesand henceacquired
'the reputationof being a moderategeneral'.87When he died in 138 the joint
governorshipwas not repeated;this suggeststhat perhapsthe greatestdanger
102
was over. The war was continuedby the new governorof PannoniaSuperior,
Ti. Haterius Nepos,who was awardedthe ornamentatriumphalia, probably also
in 138-thelast generalin Romanhistory to be so distinguished.88 His victory
was doubtlessfollowed by the usual negotiations which resulted in the in-
vestitureof a king of the Quadi. The reverseof a coin of the EmperorAntoninus
Pius belongingto the years 140/4 depictsthe investiture:the emperoris shown
handing the diademto the king and the inscription runs REX QVADIS DATVS.89
As suchrepresentations on coinsarecomparativelyrare, this one is not without
importance;it is possiblethat the negotiationswere protractedand the coin
depictedlong-awaitednews.
Peaceon the DanubeunderHadrianandAntoninusPius was only interrupted
by eachhaving to wagewar againstthe Sarmatiansand the Quadi. Recentlythe
possibility has beenfrequently consideredthat a period of unrestbeganin the
150S. The Historia Augusta agreesthat the war of 167-80 did not break out
suddenlybut had beenpostponedfor years by the governorsof the provinces
concerned.90Now there is a quite large number of coin-hoardswhose latest
coins dateto the 15 os. At first sight this suggeststhat therewas a long periodof
unrestunder Pius and in the early yearsof Marcus, in which peoplehad buried
their valuables;and that as the ownersof much of the buried treasureperished
in the confusion,it remainedin the ground. If we pursuethis idea to its logical
conclusion, we must assumefrom the distribution of these coin-hoardsthat
there were several invasions by the barbarians,who penetrateddeep into
Pannoniaand Noricum, and even as far as central Italy. This was in fact the
conclusion reached; the decade and a half preceding the outbreak of the
Marcomannicwar were interpretedas a period of increasing insecurity and
regardedas the 'preludeto the Marcomannicwars'.91
Few will supposethat the war broke out suddenly, still less that it was
unexpected.Yet repeatedsorties by the barbariansinto the Danubeprovinces
in a periodof generalpeaceunderPius would haveleft an echo.The invasionof
Italy by the Danubian Germansin the early years of the Marcomannicwars
made so deep an impressionthat not merely contemporaries,such as Lucian,
but evenAmmianusMarcellinus consideredit worth mentioning.Thereforewe
ought rather to connect the coin-hoardsending in the middle of the second
centurywith the later wars. If a personhoardsmoneyover a period of time, his
collection will not inevitably containcoins of the last few years.The circulation
of money,particularlyin remotepartsof the provincesor among thelower class
of agriculturalworkers,was probablya fairly slow process,and evenif a person
cameinto possessionof newly minted coins it is quite possiblethat he did not
hoard them. Recently it was pointed out that two hoards of coins, unearthed
103
The Danubefrontier from Vespasianto iVIarcus Aurelius
south of Carnuntum and containing terminal coins of different years, must
unquestionablybe connectedwith the samebarbarianattack: the differencewas
causedby the circumstancessurroundingthe hoarding and the income of the
owners.92The coin-hoardsof the middle of the secondcenturywere, therefore,
probably not buried until the time of Marcus, during the catastrophicyears
following 167. It shouldalso be rememberedthat moneyhadnot yet depreciated
to the sameextent as it did in the third and fourth centuries.As will be seen,
evenin the third century coin-hoardscontainingterminal coins of varied dates
canbe tracedto the samewar; this appliesevenmore to hoardscontaininggood
second-centurymoney. Peoplein later centuriescollected these coins by pre-
ference,for instanceSarmatiangravesof the third and fourth centuriesmostly
contain second-centurysilver denarii.
In the history of the Danube frontier the years of peaceunder Hadrian and
Antoninus Pius were a period of final consolidation but also one in which
inflexibility set in. As alreadyindicated,the limes-fortson the Danube,with very
few exceptions,were built between89 and 102, and at the sametime the garri-
sonsin the interior of the provincewerewithdrawn. At the presenttime it is not
possibleto describethe systemof occupationwithin the interior of Pannoniain
the first century, as the numberof suspectedor definite forts is still very small
(GyaI6ka, Mattersdorf, Ulaka, Tic-only the last two are attestedby excava-
tions, though they still have not beendated).93It is, therefore,very significant
that after the beginningof the secondcenturythereis neitherarchaeologicalnor
epigraphicevidencethat Roman troops were stationedbehind the Pannonian
limes. It follows that in contrast with the situation in Upper Moesia, all the
known auxiliary units of the exercitusPannoniaeSuperiorisand Inferioris, whether
attestedon inscriptionsor diplomas,could be accommodatedin the limes-forts.
Therewere more than thirty such forts (Fig. 18), and second-centurydiplomas
record on average 10-12 units in PannoniaSuperior and 13-18 in Inferior.
Thoselimes-fortsso far excavatedwere mannedcontinuouslyin the secondand
third centuries, and the considerablenumber of inscriptions referring to
auxiliary soldiers and veteransor their units which have been found at other
placeson the limes lead to the sameconclusion-thatthe rest of the forts were
also continuouslymanned.94
The forts on the Pannonianlimesfrom the boundaryof Noricum to the mouth
of the Savehavein the main beenlocatedand identified, but we cannotexclude
the possibility that further forts will be discovered.For a long time it was
tacitly assumedthat all of them were listed in the AntonineItinerary or marked
on the Peutingermap; field-work and excavationsboth suggestedthat those
104
places,and only those,were sites of forts. Recently,however,an auxiliary fort
of the secondor third century was discoveredand investigatedat Albertfalva
(Budapest,XI District); it was probablythe first limes-fort south of the fortress
at Aquincum, thoughCampona,the thoroughlyinvestigatedfort at Nagyteteny
(Budapest,XXII District) is listed in the Itinerary in this position. The Itinerary
doesnot mentionthe fort of Albertfalva, probably becauseit was destroyedin
the third century under Gallienus.9 5
Down-streamfrom the mouth of the Drave the sites of the forts have been
more or less pin-pointed with the help of the Itinerary and the discovery of
inscriptions,but owing to lack of field-work and evenmore so of excavationit
is by no meanscertainwhich sitesare to be regardedas limes-fortsof the second
and third centuries.On this stretchof the Danubea great deal of building was
carriedout in late antiquity; hencethoseplacesmentionedin the Itinerary were
not necessarilyfort-sites in the earlier period.
The distribution of auxiliary forts on the Pannonianlimes reflectsthe working
of the military mind, inasmuchas a basically mechanicalplan, worked out at a
desk, was adaptedto the terrain; the essenceof the plan was even distribution.
The forts were 20 km or 15 Romanmiles apart,evenin marshyareassubjectto
flooding wherea lesscompactchainwould havebeensufficient (for instanceeast
of Camuntum, or in the moderncountyof Tolna southofIntercisa).At the same
time the siteswere so chosenthat the forts could control as long a stretchof the
Danube as possible, or block the entranceto a lateral valley. The principles
involved are still somewhatobscure.In many places the loess plateauon the
limes of PannoniaInferior ends on the river-bank in a high, very steepescarp-
ment or cliff which is being underminedby the scour of the river; the steep
edge of the plateauis constantlygiving way. Despite this, some of the most
important auxiliary forts were built on the edge of the plateau. Though they
faced the river, theseforts were quite inaccessiblefrom it; moreoverthe limes-
road along the bank which connectedthem had to overcome considerable
changesof level. Becauseof the collapseof the cliff edgethe forts had sometimes
to be movedfurther back. For instance,underHadrianthe fort of Intercisawas
moved 20 m to the west, obviously becausethe front wall was underminedor
hadactually fallen into the river. Of thoseforts sitedon the plateauonly partsof
the retentura have survived (e.g. at Annamatiaor Lugio). Neitherfor offensive
nor for defensivestrategydid theseforts offer any advantage;they were, in fact,
nothing more than favourably situated observation-posts,and watch-towers
would equally well have servedthe samepurpose.Forts were only built on the
loessplateauwheretherewas a cliff edgealong the river. Whereverthe cliffs did
not rise abruptly fromthe river and there was flat groundbetweenthe slopeof
105
the plateauand the water the forts were built down below on the bank. In such
casesthey were mostly sited at the entranceto a valley or to a cutting in the
plateau,and generallyat one end of the flat land ratherthan in the middle of it.
In marshy areas subject to flooding, small island-like hills were chosen;if it
happenedto be impossibleto build the limes-roadthroughthe marshland,then a
link road to the fort was built. For instance,accordingto the Itinerary the fort at
Alisca C6cseny)lay96 ad latus-offthe limes-road.Wherethe river-bankwas not
steepand the hinterlandwas mainly flat, the forts were sited fairly close to the
bank, with the result that they were sometimeswashedaway. Only the rear face
of the auxiliary fort at Vetus Salina has survived, and in somecasesthe whole
fort has disappeared;it should also be rememberedthat Poigen, where the
station Villa Gai of the Peutingermap was situated,vanishedwithout trace in
the sixteenthcentury through erosion.97
The earliestperiodsof the forts so far investigatedhave so much in common
that a generaldescriptionis appropriate.In the past, all the forts so far located
down-streamfrom Brigetio as far asIntercisahavebeenexcavatedto a greateror
lesserextent:98 more recent excavationshave producedmany details of some
aboveBrigetio in present-dayHungary.99In all the forts excavatedby modern
techniques-and thesefortunatelyform the majority-tracesof earth-and-timber
forts were found, consisting of a ditch, rampart and post-holesand, what is
rarer, sun-driedclay bricks and the foundationsof woodenbarracks.Except at
Intercisa,wherethe superimposedstonefort doesnot coincidewith the palisad-
ed one, the remainsof the timber fort were destroyedin erectingthe buildings
of the stonefort, since the latter correspondedpreciselyin both shapeand size
with its predecessor. At Intercisaonly a singlepalisadedfort could be identified
and in other forts thereare tracesof only a single timber period below the wall
and ditch of the stonestructure.This can hardly be accidental.The earth-and-
timber forts built in the Flavian period could hardly have survivedwithout re-
building for three, four or even five decadesuntil their replacementin stone
under Hadrian or even later. The timber forts below the latter probably date
only to Trajan, and the Flavianforts (and evenearlier oneshereandthere)were
not on the samesites as the later stonestructures.This assumptionis supported
by the fact that exceptionallyat Vetus Salinaseveralforts of earlierperiodshave
been identified underneaththe stone fort. However, on other well-excavated
sitesrescuework hasrevealedtracesof earthworkssouthof the stonefort, as for
instanceat Matrica and Intercisa.1oo
This meansthat the sites of the limes-fortswere not finally fixed until Trajan.
Previously, troops and forts were much less permanentlysited. The building
activity undertakenin Trajan'sreign markedthe beginningof a new policy with
106
regardto the forts on the limes; Hadrianmerely continuedit when he had them
rebuilt in stone,or at leastwhenhe startedlarge-scalebuilding operationswhich
in someplaceswere possiblynot completedevenunderPius.
Thoughinformation concerningtypes of earth-and-timberforts in useis still
lacking, thereis a substantialamountavailableabout the plan and construction
methodsof the stoneones. Since they show no basic differencefrom the well-
known type found in Britain and on the Rhine, they must be associatedwith the
building activity startedby Hadrian (Fig. 1 S). Stratigraphicobservationscon-
firm this dating.
The stone forts built under Hadrian or in the mid second century have
roundedcorners,rectangularor, often, squaregate-towersand internal corner-
towers built to a trapezoidalplan. The thicknessof the walls is generally no
more than 1·20 m, thougho·So m was not unusual.So far hardly any buildings
within the forts have been uncovered,but one thing is certain: the defensive
walls were strengthenedinside with a rampart embankment.Constructionon
the limes was probably startedafter Hadrian'svisit in 124; at the sametime the
limes-road connecting the forts was paved with stone. North of the fort of
Matrica a fairly worn coin of Hadrian inscribedcos III (i.e. not minted before
II9) was found underneaththe thick metalling.IoI It is probablethat the limes-
road was metalledbefore the forts were built; henceit would not be wrong to
datethe forts to the middle of the secondcentury.In the stonefort at Quadrata
in PannoniaSuperior, which has recently been investigated,the rampart was
thrown up, in placesat least, a good deallater.I02
In addition to forts, smaller fortifications, mostly watch-towers,were built
during thesefirst extensiveoperationson the limes. Trajan's Column, with its
reliefs of burgi on the Danube,providesindisputableevidenceof their existence.
Theseburgi were made of wood, but so far very few have been uncovered-
thereare,for exampleonenearFischamendand othersnearDunaalmasandnear
Nyergesujfalu,all on the Upper Pannonianlimes.Io3 The one nearNyergesujfalu
was stone-built.The numberof such burgi from the early period which have so
far beeninvestigatedis so small that no generaldeductionsmay be drawn. It is
probablethat a numberof the towersandfortlets identified by surfaceinspection
on the bank of the Danubein the Djerdapby the older generationof archaeolo-
gists, particularly by E. Swoboda,dateto the earlier imperial period, sinceforts
could be built in only a few placesin the Djerdap(PI. Sa), but the cliff road had
to be constantlyguarded.104
It is probably no accidentthat the earliestmilestonesin Pannoniadate from
Hadrian and Antoninus Pius (some of the stones,with their unusual way of
indicating distance-e.g.a Malata Cusum-arememorial stonesto Nerva which
107
Danube
QVADRATA
VLClsrA
QVADRATA
100 metres
MATRICA
QVADRATA
QVADRATA
Figure IS Danubeforts in Pannonia
108
had beenrenewed).IOsThe road network in the province was obviously not
finished until the mid secondcentury, though certainly no later. An enormous
amountof labour and material was involved in the building of the limes-forts;
the stonehadfrequentlyto be transportedfrom quarrieswhich were more than
a day'sjourneyfrom the building-site.Stone,tiles, lime andwood wereprobably
carried, where possible, by water. Heavy demandswere made on the classis
Flavia Pannonica as well as on privately owned ships, but without roads con-
struction work was impossible.Moreover, this work, which probably took a
decade to complete, would have been unimaginable without some central
planning and guiding inspiration. Contemporaryand modern historians both
credit Hadrian with this inspiration, and his negative attitude towards his
predecessor's conquestsseemsto suggestthat they are right. What is striking,
however, is the fact that the plan for defending the frontier was apparently
drawn up under Trajan; Hadrian only drew from it the inevitable conclusion
andhadthe forts built of stoneas permanentstructures.If Appian of Alexandria,
writing not much later, describesthe empire as being surroundedby a closely
woven chain of fortifications which protect it 'as if it were an estate',I06and
Aelius Aristides refers explicitly to a wall round the empire/o7 then this new
conception, formulated by Trajan and put into effect by Hadrian, found
significant expression.It would perhapsbe idle to attempt to distinguish the
various stagesof this plan rangingfrom Vespasianto Hadrian.It was probably
the state of foreign policy and experienceover many years that led Trajan,
despite or perhapsprecisely becauseof his policy of conquest,to furnish the
empirewith plansfor a different systembasedon the bestpossiblefrontiers, and
Hadrianto put theseplans,with certainmodifications,into effect.
It might reasonablybe assumedthat the same kind of building work was
carried out on the short sector of the limes in Moesia Superior. There was,
however,an important differencehere in that the interior of the province was
not demilitarized. In fact the occupation army had to be reinforced under
Marcus Aurelius.
As a result of the concentrationof the Danubearmy on the limes, a process
startedby the Flaviansand completedunderTrajan,the river becameone of the
most important routes in the empire. It goes without saying that Trajan paid
particularattentionto shipping.The famousTrajanictabula of the year 100 in the
Kasan gorge confirms that the cliff road had been improved by widening
(PI. 9b); it had been hewn much deeperinto the rock-face, and the wooden
structure which had jutted out over the river was replaced by something
simpler. A recentlydiscoveredinscription of Trajanatteststhe constructionof a
canal in the Iron Gates: ob periculum cataractarumderivato ftumine tutam Danuvi
1°9
navigationemfecit. It was not until the end of the nineteenthcentury that a later
canal with the samepurposewas madein the Iron Gates.108
No useful detailsare available aboutthe harbourinstallationson the river.109
The earlier landing-placescloseto the water had probablyalreadybeenwashed
away. From certainlanding-placeswhich have survivedfrom late antiquity we
may draw the generalinferencethat the earlier ones-mainlyFlavian or Traj-
anic-werefloating jetties. The army neededthem, not only becauseof the
considerableamount of transportby water but also because,since at least the
time of Trajan,Romanstrong-pointshadbeenbuilt on the left bank of the river
which could only rarely be reachedfrom the oppositebankby a pontoonbridge.
The postsof a bridge which connectedthe legionaryfortressat Aquincum with
the bridge-headfort on the oppositebank nearthe mouth of the Rakos(Buda-
pest,XIII District) were discoveredin the river-bedin the last century.IIO They
came either from a wooden bridge built for some specific purposeand not
intendedto last longer than one summer,or from a pontoonbridge supported
on extremelystrongpiles. It is hardly likely that a woodenbridge would have
survived a winter in view of the very severeice drifting. Most of the Danube
crossingswere probably madeby ferry.
Many of the forts on the left bank of the Danubehave disappearedwithout
trace. The only one to be thoroughly excavatedwas an auxiliary fort covering
an area of 175 by 176 m, discoveredopposite the legionary fortress of Bri-
getio;III oppositeCarnuntumthe only discoverywas the remainsof a building
(the 6desKloster) which may not originally havebeensitedon that bank,II2and
oppositeAquincum the fort on the Rakos was a walled quadrilateralwith an
inner portico on three sides.I13 The fort across the river at Brigetio, IZa-
Leanyvar, can be identified with Celamantia,mentionedby Ptolemy; it is not
directly oppositethe fortress but a little down-stream;at Aquincum, too, we
may assumethat there was a fort near the ferry down-streamfrom the fortress
andoppositethe late La Teneoppidumon Mount Gellert. It washerethat Roman
hypocaustswerefound outsidethe late Romanfort. 114 Furtherfortifications have
beendiscoveredoppositethe auxiliary forts of Lussonium(DunakomlOd)and
Lugio (Dunaszekcso): lIS their date is, however, not at all certain. There are
inscriptionsfrom the time of the Severi built into the walls of the fort opposite
Lugio II6 which suggesta late date; but as Lugio was the terminal of the road
through Sarmatianterritory (attestedon an inscription from Szeged),it was
necessaryfor the crossingto be guardedon the left bank. It must, therefore,be
assumedthat someauxiliary forts had counter-fortsacrossthe river as early as
the secondcentury. It is probablethat fully developedforts were built only at
certain places,mainly oppositelegionary fortresses.The others seemto have
IlO
beensimple rectangularopenareasof about0·5 ha enclosedby walls.II7 Sigillata
pottery from the fort at IZa-Leanyvar begins with the so-called Po valley
productsof the Flavianperiod,II8 andsuggeststhat it was built at the time of the
Suebianwar before a legion was stationedat Brigetio. After the end of the
Danubianwars theseforts on the left bank probablyassumeda quite different
role. They were an organic part of the limes system;they did not serveas the
spring-boardfor advancesplannedas part of an expansionistpolicy, which in
any casecameto an end with Trajan. It goeswithout saying thatRomerealized
the advisability of continuingto demonstrateits 'military presence',not leastas
a means of supporting its client-princes.But this could have been achieved
equally well by the chain of forts on the right bank. The counter-fortsserved
chiefly to safeguardpeaceful contactswith barbarianterritory. This did not
mean that transportalways had to have a military escort, but that a prudent
system of treaties had to be guaranteed.The protection of communications
undersuchconditionswas, therefore,probablymoreof a policing task.
Nothing throws into sharperrelief the self-confidenceof the Romans,and
the peacefulconditions which were seldom disturbed,than the palaceof the
governorof PannoniaInferior which was built at the latest in the mid second
century.II9 It stood on the bank of the Danube in front of the fortress at
Aquincum; the ceremonialapartments,a suite of imposing rooms decorated
with mosaics,stuccoand murals, coincided,so to speak,exactly with the outer
line of the imperial frontier, and the main front of this imposingbuilding faced
the barbarianlands.It is impossibleto imaginethat it could havebeenconceived
in a fit of aberrationor forgetfulness on the part of the architect. It was a
consciousdemonstration,to those within as well as those without, of that
contemporaryfeeling to which, for example, Aelius Aristides gave literary
expression.Such a striking architectural style, particularly in the case of a
governor'spalace,must be attributednot to a whim but to carefully considered
political strategy:of this the governorhad to be the prime embodiment.
III
ChapterJ
The hrst age of prosperity
The interval betweenthe Year of the Four Emperorsand the outbreakof the
Marcomannicwars,lasting almosta century,can be divided into two periodsof
more or less equal length, which, despite the exigencies of foreign policy,
emergeas logically connectedstagesof a systematicplan. The latter was, of
course,only the consequenceof applying in the peacefulpost-Trajanic period
conclusionsdrawn from a policy which had developedgradually under the
Flavians and Trajan. In Pannoniathis policy was crowned with success;in
MoesiaSuperior,however,it cameup againstdifficulties which causeddevelop-
ment to be divertedto an entirely different direction. In generalthe first period,
lastingfrom Vespasianto the momentwhentroop deploymenton the Danubian
limes was complete,can beregardedas a warlike situation;in the secondperiod,
roughly from the last yearsof Trajan to the outbreakof the Marcomannicwars,
the framework thus brought into being was consolidated.
The spreadoj cities

An indication of the Flavians' new policy was the foundation of two colonies
soon after Vespasiancameto the throne. The Elder Pliny calls Sisciaa colonia,I
thoughhe doesnot as yet apply that term to Sirmium, the other colonia Flavia
in Pannonia;both towns were, however, given as the origo of Roman citizens
who were witnessesto a military diploma2 dated73. Thus a deductio must have
takenplacein Sirmium as well as Sisciain the period 70-3.3 A military diploma
was issuedin 71 to a man serving with the Ravennafleet,4 which, accordingto
the formula e! sun! deducti in Pannoniam,was treatedas an exceptionalcaseand
IIZ
The first age oj prosperity
awardedthe missiohonestathrougha deductio. The sailors attachedto the fleets at
RavennaandMisenumwere probablytreatedin this privileged mannerbecause
they had supportedVespasianin the Year of the Four Emperors.s At present
there is no evidencethat there were legionary veteransat either Siscia or Sir-
mium; on the other hand, the tombstoneof a dischargedsailor at Siscia6 sug-
geststhat the foundation of the colony at Siscia, and probably also of that at
Sirmium, resultedfrom the grant of land to dischargedmembersof the fleet.
The phrasesuntdeductiimplies that land was grantedin a colony. The sailorswho
had distinguishedthemselveson the side of the Flavianswere mainly nativesof
PannoniaandDalmatia,who on settling in the Savevalley wereprobablynot in
a position to make any significant contributionto the Romanizationof the two
new colonies. At Siscia and Sirmium the number of tombstonesof newly
settledcitizensandtheir descendants is smallerthan in Savariaand suggeststhat
the former were less intensively Romanized.The deductionsat Siscia and Sir-
mium, which were exceptionalin concerningmembersof the navy, were not
simply due to desireon the part of Vespasianto reward the classiarii who had
fought on his sideandto settlethemfor political reasonsin their homeprovince.
Part of them were in fact settledin Paestum. 7 The choice of Sisciaand Sirmium
was thereforenot accidental.There is epigraphicevidencethat the spreadof

Italian tradersinto the Save valley had been restrictedto the upper reaches.8
Hence,by settling the sailors in the two most important centresin the valley,
Vespasianprobably also wanted to encourageshipping on the river. These
former sailors, now settled in the harbour towns of the Save, were probably
given specialprivilegeswith respectto traffic on the river. In this contextit is not
without significancethat the Pannonianfleet, thoughof earlier origin, from the
time of the Flavians bore the title 'Flavia'. The possibility has already been
considered that the reorganization of the Pannonian fleet dates back to
Vespasian.
The founding of two colonieson the Saveis in someway connectedwith the
transferof legio V Alaudaeto Moesiaand with the early stagesof the construc-
tion of the fort-system on the Danubefrontier below the mouth of the Save.
This stretchof the Danubecould mosteasily be reachedfrom Italy by way of the
latter river; in developingthe supply network, the Save,and thereforealso the
dischargedmarinesat Sisciaand Sirmium, assumedfirst-class importance.
Thereis still insufficient information concerningthe topographyof Siscia and
Sirmium to make any reconstructionof the original town-plans possible.Al-
thoughextensive excavations in Sirmium in the last decadehave establishedin
the main the extent of the late Romanimperial city as well as the main streets,
many public buildings, the line of the town-wall and the site of the early Chris-
I13
tian cemeteries(Figs 29, 52), the pattern emerging is relevant only to late
antiquity, by which time the town had probably increasedseveraltimes over.9
Excavationshaveso far failed to discoverthe site of the earliertown-centre,that
of the Flavian colony. The samianpottery found then or brought to light in
earlier excavationsand now awaiting examinationin the museumat Sremska
Mitrovica is, however,in such quantity that a vigorous commerciallife and a
wealthypopulationin the first andsecondcenturiesmust be assumed.The town
probably stood a little distancefrom the Save, since the harbour, which was
enclosedby the town-walls in late antiquity, was linked then to the town by a
double wall and was clearly therefore not adjacentto it. Siscia stood at the
confluenceof the Colapis (Kulpa) where both rivers bend, so that the town,
whetheron the right or left bank of the Kulpa, had rivers on three sides.It is
still not certainwherethe Flaviancolony, or evenwherethe Pannonianfortified
town of Segestike,capturedby Octavianin 35 B.C., was situated.If we follow
Count Marsigli (1658-1730), an able military engineerwho made a plan of
Siscia,we shouldseekthe town on the left bank of the Kulpa where,according
to him, the town-walls andthe gateswere sited betweenthe Kulpa andthe Save
at Stari Sisak. On a hill on the right bank of the Kulpa and oppositeStari Sisak
is Novi Sisak, which the surfaceconfigurationwould suggestmight be the site
of Segestike,were it not for the fact that thereare still dykes round Stari Sisak
which could be those of Tiberius.IO The topography becomeseven more
obscurewhen we considerthat the searchfor Segestike,for the headquartersof
Tiberius,for the assumedlegionarycampandfinally for the Flaviancolony need
not necessarilyall involve the sameplace.II
The territories of Siscia and Sirmium were inhabitedby the Pannoniantribes
of the Colapianiand the Amantini respectively.It is probablethat their civitates
peregrinae ceasedto exist in 71 and that their tribal aristocracywas gradually
mergedin the city councils Cordo). Romancitizens who had acquiredthat status
under the Flavians are attestedfrom both towns.I2 Part of the land was, of
course, divided among the sailors. An interesting monument, a boundary
stone13-theonly one of its kind in Pannonia-wasdiscoveredat Cerevicnorth
of Sirmium. This stonewas on the boundaryof the vicusIosista, all of which had
beenassignedto an officer of the EquestrianOrder. This may be deducedfrom
the phraseadsignatusTib. Claudio Prisco prCaefecto)a/Cae) I c(ivium) R(omanorum).
If Priscusas a iudex or arbiter had assignedthe vicus lands to someoneelse (to
whom?), thenthe inscription wouldreadper Tib. CI. Priscum.Plots of land were
thereforegrantednot only to sailors but also to high-rankingofficers. From the
officer's nameit follows that it was only under Claudius or Nero that he was
grantedcitizenship; henceit is highly improbablethat as early as 71 he was
II4
already in a position to function as an ala-prefect of equestrianrank. It was
probably under the Flaviansthat he reachedthis command-thethird stageof
the militia equestris-fromwhich it may be deducedthat the allotment of land in
Sirmium did not end with a single act of deductio in 71. It is obvious that the
land allotted to a common sailor was not the same size as that grantedto an
officer. The land attachedto a vicus was, relatively speaking,extensive,and this
explainswhy the big estatesin the neighbourhoodof Sirmium, of which more
later, had their origin in the first century.
The other towns createdby the Flavian policy of urbanizationin Pannonia
were municipia, which will be consideredin anothercontext in this chapter.
Here it need only be emphasizedthat the three municipia-Neviodunum,
Andautoniaand Scarbantia-layon roadswhich in the Flavianperiod were the
most important in the province (Fig. 37). The founding of Neviodunumand
Andautoniaresultedin a chain of urban communitieson the upper Savewhich
made the founding of further towns unnecessarythere. This chain obviously
formed part of the plan, mentionedabove,for encouragingtraffic on and along
the Save. Scarbantia,which as the oppidum of ScarbantiaJulia had enjoyed a
certain corporateorganizationsince the time of Tiberius, was situatedon the
Amber Roadbetweenthe Claudiancolonia Savariaandthe legionaryfortressat
Carnuntum.The municipalizationof this existing group of citizens was prob-
ably not such an important turning-point in the history of this settlementas a
deductio would have been.Of greaterimportancewas the new policy which led
to the founding of the first municipia,threeof which now madetheir appearance
in Pannonia.The grantof municipalstatusto Scarbantiaitself could just as easily
have been made under Nero, or even Claudius. That it was given this status
underthe Flaviansmeansthat this stepsuitedtheir new policy, onewhich led to
the creationof five cities in Pannonia,more than twice as many as under the
Julio-Claudianemperors.
In Upper Moesia the founding of the first city is also attributable to the
Flavians.It is, however,significant that in this province it was not possibleat
that time to createmunicipia, as the necessaryconditions did not exist; there
were neither groups of citizens made up of foreign settlers nor any well-
established,more or less Romanized,civitatesperegrinae.The Flaviansalso pro-
ceededvery cautiouslywith their policy of urbanizationbasedon the deductio.
The only Flavian colony was in the Axius (Vardar) valley, not far from the
Macedonian-Moesian boundaryand in an areawhich was largely isolatedfrom
the rest of the province,and which evenlater on lackedmuch cultural connec-
tion with the northernparts.From the point of view of easeof communications
andland suitablefor agriculturesomeplacesin the Kosovo Polje or in the lower
IIj
The first age oj prosperiry
Morava valley would have proved to be better choicesfor a deductio than the
neighbourhoodof Scupi CSkoplje). The upper Vardar was probably chosen
becauseat that time it was regardedas being the most completelypacified area,
and the north-southline of communicationrunning throughthe Vardar valley,
the Kosovo Folje andthe Moravavalley was regarded,no doubt with reason,as
havinga certainimportance.Southof Scupiin the Vardarvalley a tombstonewas
discoveredof a veteranwho had previouslyservedin IIII Macedonica,a legion
disbandedby Vespasian:its owner must have settledthere in the pre-Flavian
period.I4 This implies a settlementnearScupi- andonewhich exerciseda certain
attraction- in pre-Flaviantimes. This settlementhad perhapsbeenestablished
close to a legionaryfortress. If this assumptionis correct, then the colony was
founded on the site of the abandonedfortress, a frequent occurrence.
From its original title DComitiana), which was soon given up, it can be
concludedthat the colony cameinto being at Scupi as the result of a deductio in
the early yearsof Domitian'sreign.Is The namecolonia Flavia D C...) has been
found on only two tombstones,which also happento be amongstthe earliestset
up by veterans. The early dropping of the epithet Domitiana can only be
explained by the damnatio memoriae which befell the founder of the colony.
Moreover,it must havebeenfoundedbeforethe outbreakof Domitian'sDacian
wars, sinceveteranswho were settledthere had formerly belongedto legio V
Alaudae,which was destroyedin the early days of the fighting. I6 Fortunatelyit
is possibleto list virtually all the legions which provided veteransfor the new
colony. Most of them camefrom VII Claudia, and somefrom V Macedonica,
I Italica andV Alaudae,in other words from legionswhich were stationedat the
time in Moesia. That VII Claudia provided most of the veterans and also,
accordingto tombstonesfrom Scupi,the earliestdecurions,is possiblydueto the
fact that it was veteransfrom this legion who were thefirst to be involved in the
deductioand were joined by thosefrom otherlegionsat a later dateunderDomi-
tian, or evenunderTrajan. The term deductusor deducticiusoccursonly on tomb-
stonesset up to veteransfrom VII Claudia.I7
The colony was not within the areaof modernSkoplje but on a hill a little to
the west at the confluence of the Vardar and Lepenacwithin the district of
Zlokucaneand Bardovce.Apart from a few housesonly the theatrehas been
investigated,rBand thereare still no details of the different periodsof the street-
grid. It was only after the disastrousearthquakeof 518 that a new town was
built and fortified under Justinian on the site of present-daySkoplje. Many
tombstoneswere discoveredin the cemeteriesat Bardovceand Zlokueane,but
two-thirds of thoseset up to veteranswere found north and southof the town;
and it is from this distribution that certain inferencesmay be drawn. The first
II6
significant detail is that only one of the veteransburied at the colony was not ~
decurion,and of thosewho were, only two were buried outsidethe town. The
earliestmissionesagrariae were probably establishednorth of Skoplje and eastof
Lepenacwhere,with one exception,all the tombstonesare thoseof veteransof
legio VII Claudia, who were either decurions or describedas deducticii. This
latter term is found only on tombstonesin this area.Later allocationsof land to
other legions are attestedwest of Lepenacand south of the Vardar; discharged
praetoriansand auxiliary veteranssettledin the southernpart of the area.In the
easternhalf of the territory, north of the Vardar,thereis no evidenceof veterans,
although there are just as many tombstonesthere. The veterans'plots were
roundthe town within a radiusof about20 km, andtheir owners wereburiedon
their own land (in praedio suo accordingto one tombstone).I9This customwas
also later continuedby the urbanaristocracy.This classwas descendedfrom the
veterans;the decurions and the wealthy citizens who were buried on their
estatesround Scupi were the descendantsof the veteranswho had beensettled
at the end of the first century.It is significant that eastof Scupi,wherethereis no
evidenceof veterans,decurionsof a later period also left no inscriptions.On the
other hand there is evidenceof someAugustales,20who probably were able to
acquireland only there.
The veteranswere an extremelymixed group. They probablyincluded more
citizens of the provinces than Italians. Apart from one from Placentia,only
Dalmatians (Salona), Macedonians (Beroia, Stobi), southern Gauls (Lucus
Vocontiorum) and Syrians (Berytos) are attested.21 In view of the method of
recruitmentto the legionsconcerned,it may also be assumedthat Noricans,men
from Asia Minor and Spaniardswere represented.This is of particular import-
ance,sincein a provincesuchas MoesiaSuperior,wherethe Italianswerenot so
numerousand had arrived there later than they had on the Adriatic coastor in
westernPannonia,the heterogeneous natureof the population was such that it
was possiblefor Latin to becomethe mother tongue of certain of the towns.
Syrians, Gauls, Macedoniansand Italians could only communicatewith each
other in Latin, which they had all learnt during their long service with the
legions.This was the reasonwhy the new colonia Flavia on the southernborder
of Upper Moesia, and very close to GreekMacedonia,could becomea Roman
town.
All that was possiblein the circumstanceswas probably achievedunder the
Flavians in the way of new municipia and coloniae. The founding of towns in
Pannoniaalong the two main arteriesof trade and troop-movement,the amber
road and the Save valley route, was virtually complete. Such towns as were
establishedlater on theseroadscould hardly be regardedas towns in the strictest
II7
The first age of prosperity
senseof the word. In Moesia conditions conduciveto the founding of towns
were muchlessfavourable.Under Trajanthe situationhardly changed,or, more
accurately,the processstartedby the Flavianscontinued;Hadrianwas the first
to graspits administrativeimplications. Trajan'spolicy for coloniae in Pannonia
and Moesia was in part a continuationof that initiated by the Flavians and in
part was a reflection of military policy, and hencewas only indirectly connected
with local conditions.If it is assumedthat under Trajan the numberof legions
in eachprovincial army correspondedto that of the coloniesin the province,
andif it is bornein mind that the permanentnumberof legionsin eachprovince
was to all intentsandpurposesfixed by him, then the conclusionfollows that he
createdthe new colonies(Poetovioin Pannoniaand Ratiariain Moesia) as new
recruitmentareasfor the provincial armies.
Poetovio was a legionary fortress throughoutthe whole of the first century,
and was only evacuatedshortly before the establishmentof the colony took
place.It is not impossiblethat up to the time of Trajantherewas also a legionary
fortress at Ratiaria. In Upper Moesia,which had beenpacified to a very limited
extent, the choice of site for the new colonia Ulpia possibly restedon the fact
that the vicinity, as a former military district, was regardedas pacified. An
additional considerationwas possibly their strategic commercial positions:
Poetovio, at the point where the Amber Road crossedthe Drave, was also
connectedwith the Danubeby a route down the Drave valley and with Aquin-
cum by a road striking off to the north-east.Ratiaria lay at the spot where the
shortestroute connectingthe Adriatic with Dacia reachedthe Danube.
In Poetoviothere are inscriptions to veteransof legionesI and II Adiutrix,
and possibly also of IIII Flavia, who becamecitizens of the new town on the
basisof severalmissionesagrariae and missionesnummariae.22 The town also yields
positive proof that the deductionof a colony was not necessarilythe result of a
single mass settlementof veteranssince, according to the inscription on his
tombstone,a veteranwas grantedland in Poetovio on the basis of a second
missio agraria.23 As well as missionesagrariae there is also specific mention of a
missio nummaria.24 The colony was probably founded on the site of the aban-
donedlegionaryfortresswhich lay on the southbankof the Dravein the areaof
present-dayZgornjaand SpodnjaHajdina nearPtuj. But a bridge was also built
over the river and a new part of the town soonsprangup on the north bank.25
Ratiaria'sveteranscamefrom legio VII Claudia, but only a very few can be
identified.26 This is also true of the veteransat Poetovio.The majority of those
for whom there is epigraphicevidenceat Ratiaria were also decurions.Their
inscriptionsare concentratedin the town. If this situationis comparedwith that
revealedby the epigraphicmaterialat the Claudiancolony at Savariaand at the
lI8
Flaviancolony at Scupi,wherelegionarieswere likewisesettledandwheremany
moreinscriptionshavebeenfound-notconfinedto the town-weareforcedto
the conclusionthat there had been a changein the policy of deductio. Perhaps
urban developmentwhich had alreadybegunin the vicinity of the abandoned
fortress formed the basis of a city. Epigraphic and archaeologicalmaterial
establishpretty clearly that this was so at Poetovio.27 If Ratiaria was a Flavian
legionaryfortress,a civilian settlementnearby mustalsobe postulated.Thus the
nucleusof a town alreadyexisted,and the few veteranswho were grantedland
in the new colony there were only the aristocratsamong a town population
alreadyin being.
Trajan's veteran settlementswere the last examplesof city foundation by
deductio in the history of the empire. Throughout his reign the founding of
colonies was still linked to the allotment of land to veterans. Poetovio is
referredto in a note on Trajanic land-allotmentin Pannoniaby the gromatist
Hyginus.28 On the other hand it is uncertainwhetherthe Hadrianic colony at
Mursawas basedon a deductio.29 This colony was the last to be establishedin the
Danubearea,possiblyin the empire,as a fresh autonomouscity. Later colonies
started as municipia and were given the title of colony under the Severi or
occasionallylater. But the colonia Aelia Mursa was not basedon an earlier
municipium. Hadrian'swell-known remark, quotedby Aulus Gellius,30 to the
effect that he was surprisedthat a city which could enjoy its own customsand
laws as a municipium shouldwish to changeits statusto that of a colonia, would
suggestthat this was unlikely. Moreover,our sourcesare unanimousthat Mursa
was foundedby Hadrian. It is not known whetherveteranswere settledthere,
nor does the still obscuretopographyof the town provide any information on
this pointY
Developmentoj trade
The founding of towns was thus basedeither on the settlementof veteransor
on existing settlementsof Roman citizens who had come in from outsidethe
province. The Flavian-Trajanic periodwas one of massiveimmigrations,but it
was not until Hadrian that they formed the nucleusof an urbanizationpolicy.
Without exceptionall the towns founded both by the Flaviansand by Trajan
weresitedon the two roadswhich at just that time werelosing their importance
to the Danube road. By contrast, the immigration of the Flavian-Trajanic
period was affecting the frontier zone where urbanizationhad not yet begun.
In fact one should speakof a secondimmigration into Pannonia.The first,
alreadytouchedupon in Chapter3, involved Italian tradersfrom northernItaly,
119
mostly from Aquileia, who took over the newly establishedmarkets on the
Amber Roadand in the upper Savevalley and probably also the externaltrade
and its chief entrepotat Carnuntum.Although this network of trade was in its
heydayat the beginningof the Flavian period, it was still by no meansequalto
the new demandsmade upon it. The very moderateneeds of the partially
Romanizedsoldiersin the few auxiliary units stationedon the Danubecould be
met either from local native productionor from army supplies.It would seem
characteristicof the pre-FlavianDanubearmy that it was suppliedwith imported
goods only to a very limited extent. Whereasthe Julio-Claudianforts on the
Rhine can be dated on the basis of Arretine pottery and other finds of luxury
goods, theseare completely absentin the Danubeforts, and even in areasin
west Pannoniamore exposedto Italian trade they have not been found in
appreciablequantity. Then, under Domitian, units suddenlyappeared,whose
memberseither came from provinces in which trade was more advanced,or
were so numerousthat they were able to createa new situation on the Danube.
Instead of a few cohorts or alae, each composedof some 500 or 1,000 half-
barbariansoldiers,in a very short spaceof time severallegions and an auxiliary
army many times its original size had to be reckonedwith, and all this in an area
which had not previouslybeenfaced with a supply problem of this magnitude.
It is obvious that the first to operatein the new marketswhich had comeinto
being as the result of the establishmentof new forts were the Italians, that is to
say the representativesof the Aquileian trading houseswhich alreadypossessed
branchesin west Pannonia.Analysis of nameson tombstonesfrom the end of
the first and the first half of the secondcenturiesoccasionallyprovidesdetailed
information about how this network of trade arose,32The families of the Canii
and Caeserniimay be cited as examplesof thoseinvolved.33 Both were originally
from Aquileia but had settledduring the first centuryin Pannonia,the Caesernii
at the beginningof the centuryin Emonaand the Canii later on in Savaria.From
the Flavian period onwards we find further establishmentsalong the diagonal
road leadinginto the interior of the provinceand ultimately to the Danube;the
Caeserniifinally settledin Brigetio, Aquincumand Sopianaeandat the mouth of
the Sala(Zala) not far from Lake Pelso;the Canii also settlednearthis lake and
at Aquincum (Fig. 19). At the southerncornerof the lake the only goodpoint at
which to crossthe Sala'smarshyriver-mouthwas nearZalavarandFenekpuszta,
wheretombstonesand a rich cremationcemeteryindicatethat there was a very
early Romansettlement.Thesesimple tombstonesrecord not only Tiberii Iulii
but also the Opponii, anotherfamily from Aquileia.34 The diagonalroad from
Savariato Mursa passedthrough Fenekpuszta, and its most important trading
post beforeMursa was Sopianae.Evidencethat the Caeserniithere camefrom
120
KAKN
AlN
I lI
CAESERNU
KANlI KANlI
MARCIl
o 50 100 ·km
OPPONII
Figure I9 The distribution of someItalian families in Pannonia
the Savariabranchof the family is providedby their relatives,the Marcii, who

are frequently recordedon inscriptions in Savariaand on tombstonesboth at
Zalavar, not far from Fenekpuszta,and at Sopianae,as being related to the
Caesernii.35Around the turn of the century there is epigraphic evidenceat
Aquincum of one C(a)esern(i)usZosimusnatione Cilix who obviously at one time
hadbeena Cilician slaveof the Caesernii,36They reachedBrigetio by way of the
short diagonalroad along the river Arabo (Riba) and were not the only Italian
family who found their way via Savariato this fortress.In addition to Q. Anto-
niusQ. j. Cl. Firmus Savaria37 it is probablethat other civilians left Savariato
121
settle in Brigetio in the first half of the secondcentury.38The Canii also came
from Savaria. There is evidenceof them north and south-westof Lake Pelso
and at Aquincum, which suggeststhat they took the other important diagonal
road which ran from Poetovionorth of Pelsoas far as Aquincum and crossed
the Savaria-Mursaroad near Fenekpuszta.
It is impossibleto list here the namesof the numerousItalian families who
settled at Brigetio, Aquincum and other places of lesser importance on the
Danubeat the end of the first century. It is, however,probablyno coincidence
that the distribution of late Italian sigillata pottery (Fig. 20) and the so-calledPo
valley warescoincidedwith the migration of the Caesernii,Marcii and Canii, to
cite only them as examples.Po valley pottery (of forms Drag. I, 2 and 25) is
extremelyrare in the interior of the province but exampleshave beenfound in
many placesnear Lake Pelso,as for instancein a cemeterynot far from Fenek-
puszta(at Keszthely-Ujmajor)and at Sagvar,a settlementat the junction of the
road south of the lake with a minor one from Mursa to Brigetio. The sameis
true of the distribution of late Italian sigillata bowls with barbotinedecoration.
The most recentdistribution map clearlyindicatesthe importanceof the Amber
Roadand the roadsfrom Savariato Brigetio andfrom Poetovioto Aquincum.39
In a variety of ways archaeologicalmaterialrevealsthe direct links connecting
the west Pannoniantowns with Aquincum or Brigetio. The tombstonereliefs at
Brigetio are probably derivedfrom thoseof Savaria,40with the possibility that
stonemasonsmoved from there to Brigetio at the beginning of the second
century.It is typical that in producingsuchsimple objectsas earthenwarelamps
the craftsmenat Aquincum copiedthe productsof Poetovio.41 Under Domitian
the demandfor industrial goodsprobablyincreasedby leaps andboundsin the
Danubeforts, with the result that a brisk traffic developedon the roads which
linked themwith the industrially more advancedtowns in west Pannonia.In the
secondstagewest Pannonian-Italianentrepreneurssettledon the Danube;this
meantthat it was no longer necessaryto transportgoodsover long distancesas
they were made on the spot. Excavatedmaterial from the Danubeareaof the
period from Domitian to roughly the reign of Hadrian contains a wealth of
imported goodsfound neither before nor afterwards.
Only part of these imported goods came from Italian or west Pannonian
workshops.As soon as there was a good numberof garrisonsalong the whole
length of the Danube,commercialshippingbeganto flourish and long-distance
river-borne trade soon proved superior to that along the roads. Industrial
centresin Gaul and Germanysooncameto recognizethat their goodscould be
carried safely and cheaply on the Danube to Pannoniaand even to Moesia.
South Gaulish manufacturershad tried earlier on to dispatch their goods to
122
date date date

date
KAN
lI
KANlI
KAN
lI
Figure 20 Italian terra sigillata in Pannonia
Pannoniavia Aquileia, so that at a fairly early datetheir samianbeganto appear

in the southandwestof the province.The Danube,however,provideda unique
opportunityto extendtheir markets.Around the turn of the centurytherewas a
flourishing Gallic-Germantrade along the river, and soon it succeededin
ousting that of Italy whereverthe great distancesoverland resultedin goods
from the west becomingcompetitive.
It is impossibleto overratethe importanceof the resultsof this changein the
pattern of trade. To begin with, the reverse side of the picture should be
considered.The areaaroundLake Pelsoat the junction of the two mostimport-
ant diagonal roads (Savaria-Mursaand Poetovio-Aquincum)was excluded
from the increasein traffic, as it was, of course,no longernecessaryto transport
imported goods overland over long distances,from west Pannoniato the
123
Thefirst age oj prosperity
Danube. The decline both in economiclife and in cultural standardsis fairly
clear in the areaaround Lake Pelso. The tombstonesfound north of the lake,
which werethe work after all of craftsmenin Poetovioand Savaria,datewithout
exceptionto the first half of the secondcentury.42The cemeteriesshow signs of
becomingpoorer,andin placeof the generalprosperitywhich was beginningon
the Danubeat this time we find that only large villas, including here and there
someluxurious ones, were being built; thesewere no doubt put up either for
rich owners of estatesor for their bailiffs. The best known is that at Balaca,
north-eastof the lake, which is decoratedwith murals of very high quality in the
earliestpost-Pompeianstyle.43 On the otherhand,thereis no evidenceat all that
the native population sharedin this opulence,although this heyday of villa-
building north of Pelsooccurredat the very time when the nativesof north and
eastPannoniawerebeginningto setup tombstonesandthe aristocracytherewas
introducingthe expensivecustomof wagon-burials.This declinein the centreof
the provinceis all the moresignificantif oneconsidersthat the nativepopulation
particularly in this areahad begunat a fairly early date to acquiresomeof the
trappingsof Romanlife. In a native cemeteryat Cserszegtomajnorth of Fenek-
puszta a two-handledvessel of late La Tene type (PI. I b) was found bearing
the burnishedinscription da bibere.44
The ownersof thesevillas in the Pelsoareawere probablyItalians, who had
settledthereand acquiredestatesduring the Flavian-Trajanic period when eco-
nomic conditions were favourable. It is possiblethat thesewere managedby
servile bailiffs (vilici). L. PetroniusL. lib. Licco, who married a native woman,
Galla Cnodavij., may havebeena freedmanwho actedin this capacity.45Whether
he was identical with a freedmanof the same name who appearson a list of
liberti at Aquileia 46 is not certain, but by no meansimprobable.Another com-
ponentof the wealthy upperclassin the areawas the veterans,who at the time
when economicconditionswere favourablehad showna preferencefor settling
along the diagonal roads.47
Once westerntrade dominatedthe Danubeit was not only Gallic-German
goodswhich found their way to Pannoniabut their manufacturerstoo. Although
they secureda strongfoothold in towns suchas Carnuntum,Savariaand Brige-
tio,48 which traditionally belongedto the Italian sphereof influence, it was
primarily in Aquincum that they outnumberedthe Italians. Here the contri-
butory factor was that the legions transferredto PannoniaunderDomitian and
Trajan came in the main from the Rhine army; and in particular X Gemina,
which was stationedin Aquincum for a longish time, came from the lower
Rhine. At an evenearlier datefunerary sculpturesat Aquincum revealevidence
of Rhenishinfluences49 (suchas military tombstoneswith full-length figures, or
124
DHOCSE
PYLCRV
C NTVLER
CI ESAG I
PINENS L X
II
Figure 2I Inscription set up by the burial-club of the cives Agrippinenses
eagleswith outspreadwings (PI. IOb)). Rhenishornamentthen appearson the

tombstonesof the Rhenishimmigrants and beautifully illustratesthe powerful
influence suddenlyintroducedfrom the Rhineland.5o
As early as the time of Trajan the Rhinelandersat Aquincum had formed an
organizationcalled cives AgrippinensesTransalpini, which carried out the func-
tions (Fig. 21) of a collegiumjuneraticium.51 The membersof the many collegia
which had beenestablishedat Aquincum at the beginningof the secondcentury
were foreignerswho had settledtherewhen tradewas flourishing. The collegium
jabrum et centonariorumsaw to the burial of CaeserniusZosimus,52 and also made
itself responsiblefor the funerals of the Gaul C. SecconiusPaternus53 and the
Thracianfreedman L. ValeriusSeutesdomoBessus.54 A collegium,significantly called
collegium negotiantium,is mentionedon an altar to Juppiter Optimus Maximus,
Neptuneand Mars, thereby indicating that the prosperity of thesetraderswas
attributable,in addition of courseto the chief god of the empire, to shipping
and the military.55
12 5
Apart from the collegia, the foreigners who had congregatedat Brigetio and
Aquincum also had an officially regulatedorganization.That in Aquincum was
called veterani et cives Romani consistentesad legionem II adiutricem,56 whereasin
Brigetio only the shortenedform cives Romaniis attested.57It is odd that despite
the fact that there are many inscriptions from the early period in Carnuntum,
wheretherewas alreadya civilian settlementin the first half of the first century,
so far no similar organizationis attested.The corporateorganizationof cives
Romanialsohada council consistingof decurionsandheadedby two magistrates.
Initially at Brigetio the over-all control was in the hands of a curator civium
Romanorum.From the legal point of view thesecorporationswerenot unlike the
numerous cives Romani consistentesof the late republican and early imperial
periods,which had beenfoundedby the Romantraderspermanentlydomiciled
abroad with a view to establishingautonomouscontrol over their interests.
Corporationsof this type, too, probably existedin Pannonia;their late emer-
gence,however,and the fact that they persisteduntil the time of the Severi is
attributable to quite different causes.Civil settlementswhich are often sur-
prisingly large developednear the legionary fortresses,surroundingthe camp
mostly on two sides though occasionallyon three. Traders, shopkeepersand
craftsmen-and,of course,the manyveteranswho plannedto investtheir money
from the missionummariain a businessnear the fortress-settledin areaswhich
were outsidethe jurisdiction of a self-governingcity or of a civitasperegrina, but
were under military control. Thus the Roman citizens living round the camps
were neitherincolae of a Romancommunitynor privilegedforeignersin a civitas
peregrina.However,as soonas sucha settlementhadreacheda certainsize which
could no longerbe administeredwithout somekind of organization,a corporate
form had to be devised.The original name cives Romaniconsistentesad legionem,
frequently attested in other provinces also, is a clear and legally accurate
descriptionof the situation.
There is evidenceof the expansionof thesecivil settlementsround all four
legionary fortressesin Pannonia(Fig. 22), but only those at Carnuntumand
Aquincum have been investigatedto any extent (Fig. 23).58 While there are
clear indications that they had sewers and a water supply and many other
adjunctsindispensableto urbanlife in antiquity, rectangularstreet-gridsare not
typical and town-walls were not a featureof the periodof the principate.Behind
the fortressat Carnuntumand not far from the porta decumanathere was a rect-
angular market-placeenclosedby a portico.59 Inside it severalwells were dis-
coveredwhich suggestthat its primary purposewas to serveas a cattle-market.
Tradein the Danuberegion is known to haveconsistedlargely of the import of
cattle and other foodstuffs from the barbarianareas.It is highly probablethat
IZ6
M
M M
M
M
M
VINDOBONA CARNVNTVM
M M
M M
M M M M M
BRIGETIO VIMINACIVM
M M
M
M
M
M
S'INGIDVNVM VIMINACIVM
Figure 22 Legionary£ortresses:can baea and municipium
127
The first age ofprosperity
VIMINACIVM
M
IMINACIV
V
VIMINACIVM
Canabae
Canabae
VIMINACIVM
VIM
INA
CIV
VIMINACIVM M
Canabae
Canabae
VIMINACIVM
Canabae Canabae
Canabae
Figure 2} Canabaelegionis: Carnuntumand Aquincum
128
this externaltradewas organizedto take placein the vicinity of the fortressand
under the control of the legion. In this way the fortressesbecameincreasingly
centresof trade.They were a sourceof attractionfor all kinds of manufacturers,
and hencemadea considerablecontributionto the developmentof many of the
settlementson the Danube.
Oneof the mostimportantelementsin this externaltradewas slaves;dealings
in them may well have been encouragedby the wars under Domitian and
Trajan. The considerablenumberof slavesin Pannoniawho camefrom beyond
the frontier are mainly recordedin inscriptionsdating to the Flavian-Hadrianic
period and mostly found near Carnuntumand Aquincum; as exampleswe may
quotePeregrinussutor caligarius natione Dacus, NaeviusPrimigeniusNaristus, Vibius
LogusHermundurus,Tudrus, Strubilo, etc. near Carnuntum,and Scorilo Ressatilib.
from Aquincum.6o It is strangethat so far there is no evidenceof Sarmatian
slaves,but then there was hardly any Romanexport trade to that areauntil the
time of Marcus.
It was only indirectly and belatedlythat Upper Moesia beganto sharein the
economicprosperity of the Flavian-Trajanic period. The reasonsfor this are
more fundamentalthan the simple explanationthat this province was further
away from Italy and the Rhinelandthan Pannoniawas. The large-scaletroop
concentrationsduring the Dacian wars of Domitian and Trajan probably
attractedarmy contractorsand the usual hangers-onto Moesia, but war condi-
tions, not to mention the two seriousdefeatsunder Domitian, were not con-
ducive to the establishmentof a permanentoutlet for goods. It is also very
questionablewhetherany long-termtradeon the frontier which facedthe hostile
Dacianscould be dependedupon. In the early Flavianperiodtherehadof course
beentraffic in goodson the Danubebelow the confluenceof the Save,and this
was responsiblefor the sporadicappearanceof Po valley sigillata pottery;61 but
everythingwhich atteststhe settlementof foreign manufacturersand the estab-
lishment of civilian settlementsnear the legionary fortressesdatesto the time
after the Dacian wars, and this is probably no accident.If the great wealth of
epigraphic material from Carnuntumfrom throughout the whole of the first
century is comparedwith the completelack of inscriptions from Viminacium,
wherea legion was stationedfrom at leastthe middle of the first century,thenit
must be admittedthat conditions of military life in pre-Trajanic Upper Moesia
were completely different from those in Pannonia.The earliest tombstonesat
Viminacium date to the beginningof the secondcentury. Before that probably
no one thought of setting up tombstonesin the legionary cemetery, simply
becausetherearelikely to have beenvery few relativespermanentlysettlednear
the camp; henceit was not possibleto establisha civilian settlement,particu-
12 9
larly as therewas little scopefor traders.The founding of coloniesat Sisciaand
Sirmium, clearly connectedwith Vespasian'sattemptsto encourageshippingon
the Save,is an indication that evenin the relatively peacefulperiod before the
Dacian war artificial methodswere necessaryto give a fillip to traffic on the
Danubebelow the mouth of the Save.
There was a suddenand fundamentalchangein the situation once Dacia
becamea province. Tombstonesbeganto be set up-a definite sign that com-
munity life was being consolidatedalong Romanlines62-anddemandsoonled
to the openingof a stonemason's yard at Viminacium. The earliestandstill very
simple stoneswere erectedto high-rankinglegionary officers; but before long
they were becomingelaborate,occasionallyby no meansinartistic (PI. 6b), and
both in compositionand style reflectedthe work of the southPannonianstone-
masons.63 Both at SingidunumandViminacium therearetombstonesto families
which originally came from Aquileia, for example that of L. Barbius f. fib.
Nymphodotus. 64 The stretch of the Danubefrom Singidunumto Viminacium
was vf importancein west-easttrade, particularly in trade betweenItaly and

Dacia, which was carried along the Save and Danube as far as Dierna and
Drobeta. Tradersfrom Aquileia were as much involved in this as were those
from the Rhineland. A married couple, in Aquileia consistens,had an altar
erectedin Singidunum65 and a negotiator Daciscusfrom Cologne was buried in
Aquileia.66 The questionariseswhy the Colognetradersfound it necessaryto
make a detour via Aquileia in order to trade with Dacia, instead of going via
Pannonia.Aquileia's central role was perhapsthe deciding factor: Gallic and
Germangoodswere also on salein that town.
The part playedby north Italians and Rhinelandersin the economicdevelop-
ment of the Upper Moesiansectionof the Danubeareawas neither substantial
nor exclusive.At the sametime as they were making their appearanceat Singi-
dunum and Viminacium there was an appreciableinflux from the easternpro-
vinces, and the considerablenumber of oriental colonists were putting their
stampon urbanlife in newly conqueredDacia. Theseeasterners first appearedat
Singidunum,Viminacium andRatiariain the first half of the secondcenturyand
both in numberandin influencewerein no way inferior to thosefrom the west.
As in Pannoniait was the military from the Rhinelandwho broughtthe Rhenish
civilians in their wake,so in Moesiathe appearance of thosefrom the eastcannot
be separatedfrom the military. Soldiersfrom the easternprovincesandfrom the
Greek provincesof the Balkans regularly servedin the Upper Moesianarmy,
both in the legions and in auxiliary units.67
The settlementsin Upper Moesia developedalong the samelines as thosein
Pannorua.At Viminacium a largecivil settlementdevelopedroundthefortress68
13°
and a smaller but definitely attestedone round that at Singidunum(Fig. 22).69
Inscriptions have so far failed to throw light on how the communitieswere
organized.CivesRomaniwith a curator at their headare attestedfrom Margumat
the mouthof the Morava.70 This place,which hadan importancefor the internal
trade of the Balkans,probably underwentthe samedevelopmentin the second
centuryas the settlementsround the legionaryfortresses.
The mines
From an imperial viewpoint the economicdevelopmenton the Danubewas
probablylessimportantthan theactivity begunby Trajanin UpperMoesiaand
lastinguntil Marcus,which aimedat exploitingthevery rich mineralresourcesof
the province.The economicdevelopmentof many Danubesettlementswas also
encouraged bythe exploitation of the mines, as the products of thesemainly
remoteworkings hadto be transportedto the Danubealongits tributaries(Ibar,
Toplica, Morava, Pek, Mlava, Timok). The civesRomaniat Margumwere prob-
ably engagedin the transportof raw metalsalong the Morava.
The openingup of the minesin UpperMoesiaalso led to an influx of popula-
tion, thoughthis did not havethe samefar-reachingeffects as that of the north
Italians and Rhinelandersinto Pannoniaand into the Danubetowns of Upper
Moesia.The socialstatusof the miners,it would seem,was not as high as that of
the manufacturerswho had establishedtheir business in the prosperity of
Flavian-Trajanic times. Also the managementof the mines was in the handsof
imperial authorities (procuratorsand freedmen) whose official position gave
them little interestin the establishmentof communitylife along urban lines.
The main source of information concerning the opening of the Upper
Moesianmines comesfrom the so-calledmine coins, small bronzecoins of the
secondcentury with the namesof Danubianmines on the reverse(PI. 13b).7I
Thesemineslay in Noricum, Pannonia,Dalmatiaand Upper Moesia, and prac-
tically all of them have beenlocated.The Upper Moesianmines mentionedon
the coins weremetalla Ulpiana, obviouslythoseeastofPristinain the present-day
Kosovo Polje (PI. 14a); metalla Dardanica, probably those in the Kopaonik
mountainseastof the river Ibar and to the north of the Kosovo Polje; metalla
Aeliana Pincensia,called after the river Pincus(Pek) and henceidentical with the
north-eastSerbianErzgebirge;and the metalla Aureliana, probably connected
with the place-nameAuriliana mentionedby Procopius,72andhenceto be identi-
fied with the minesroundBor in north-eastSerbia.Identifying theseminesis not
as simpleas it looks; a little thought,however,will perhapsshowthat to identify
them in any other way is not possible.Moreover,all of them must be regarded
131
AVREVS -.
MONS
XVLPIANA
Ko
Kosm~
sm
~
'-AELlANA
PINCENSIA
.AVRELIANA~
XVLPIANA
Kos
mK ~
osm
m~
~
Kos
XVLPIANA ~DARDANICA
XVLPIANA XVLPIANA
METAllA
MVNICIPIVM
0
.& 50
~'
100
km
Figure 24 Mines and municipia in Upper Moesia

as identical with the best-knownmines still operatingtoday; and it is obviously
not fortuitous that a municipium Ulpianum and a municipium Dard (...) are
known to have existednot far off. Ulpianum was south of Pristina and to the
westof Mount Zegovacand Dard (...) in the Ibar valley, westof Kopaonik. At
a later datethe nameUlpianumbecameUlpiana; this suggeststhat the Aureliana
listed by Procopiusin the xwpa (chora) of Aquae amongthe place-namesin the
Timok valley was perhapsoriginally Aurelianum, and hencethat there was a
municipiumthere.This is a justifiable inference,sinceit is known that therewas
a place called Municipium,73 probably municipium Aelianum, not far from the
mines which must be identified with the metalla Aeliana Pincensia.It will be seen
later that in the course of the second and third centuries municipia were
establishednear all thesemines (Fig. 24), and that each took the name of its
respectivemetalla (pp. 223 ff).
The Upper Moesian mines producedprimarily lead and from it silver; but
probably copperand here and there other minerals were also extracted.74 The
output was very high, and in someareasmining was very intensive. Near the
mineson Mount Kosmaj southof Singidunum, whichfirst went into production
underMarcusAurelius andfor that reasonarenot mentionedon the earliermine
coins,it was estimatedlast centurythat the tips containeda million tons of waste
material.7s It wasfurther establishedthat the oreshadbeenextractedright down
to ground-waterlevel.
The mine coins start with Trajan and mention metalla Ulpiana in several
Danube provinces. Hence the name Ulpianum originated under him, a fact
which is otherwiseestablished;it was also under Trajan that therewas the first
mentionof the Dardanica(Pl. 13b). Thus he was responsiblefor the openingand
organizationof the south Moesian mines. Somewhatlater under Hadrian the
Aeliana and possibly also the Aureliana came into production. The former,
which derive their name from him, are mentioned on coins minted under
Hadrianbut referenceto the latter occursonly on undatedmintings. The name
Aureliana may perhapsbe connectedwith the young CaesarMarcus Aurelius
and may therefore also be dated to the year previous to Hadrian's death or
possiblya little later. Whetherit was for securityreasonsthat the openingof the
minesin the southprecededthosein the north, which werecloserto the frontier,
canonly be a matterof conjecture.It wasMarcusAurelius who hadthe mineson
Mount Kosmaj nearestthe frontier started,and he had to protect them with
auxiliary units.
Unfortunatelythereis very little information aboutthe way thesemineswere
managed.The one certainfact is that originally they were in the handsof pro-
curators(procuratoresmetalIi) eachof whom was probably responsibleonly for
133
one territorium metalli and as a rule was an imperial freedman.76 Freedmenof the
emperoralso provided the technicalexpertsas, for example,P. Aelius P. Lib.
MenandercenturioofIicinarum, or P. AeliusPfato mensor.77The mineswere,however,
leasedto coloni who also had a kind of autonomy.On an inscription from the
metaffaDardanica occursthe formula focus datusdecreto cofonorum,which is remin-
iscent of focus datus decreto decurionum,78the form current in municipal districts.
Under Hadrian the cofoni argentariarum erecteda shrine to Antinous after his
deification; this perhapsstoodin the middle of the store-houses (horrea) (PI. 2oa)
belongingto the mine administration. It is probablethat the coloni brought
79
considerablenumbersof miners from Thraceand Dalmatia into the province,

and perhapsalso someslaves.Thesethen formed what was tantamountto an
urbanpopulationin the mining settlements.The upperclassin thesesettlements
was representedby the richer coloni, the majority of whom were probably
freedmen. While there is some referenceto them on inscriptions from the
metalfa Ulpiana and Dardanica, the strongestevidencecomes from the mining
district on Mount Kosmaj whosegreatestperiod of prosperityoccurredunder
the Severi.
Civitatesand municipia
Despite extensivecolonization and large-scaleimmigration, foreigners were
not the only onesto be involved in the new Flavian-Trajanicpolicy. Therewas
also a fundamentallynew approachto the native population,despiteconsider-
able differencesbetweenPannoniaand Upper Moesia. It was at this point that
the developmentof the two provincesdiverged.
The civitatesperegrinaeweregrantedautonomyeither under the Flavians or at
latestunderTrajan, or to be more precisethey were freed from military control
and were placedunderpraeJectichosenfrom the ranks of the native aristocracy
(principes). Naturally it is not certain just when this changetook place,and it is
evenless certainwhetherthis form of local self-governmentwas grantedsimul-
taneouslyto all civitates.Possiblysomenorth Pannoniancivitateswerenot freed
from control by military praeJectiuntil the time of Trajan.It would seemthat the
introduction of local praeJecti was connectedwith fairly widespreadgrants of
citizenship;but grantsunderthe Flaviansare attestedin only a few civitates in
the southandwestofPannonia,andin someareasthey beganonly underTrajan.
For a long time analogiesin Dalmatia had suggestedthat the military praeJec-
turae were abolishedtowardsthe end of the first centuryand replacedby local
functionaries.It was not until 1956 that positive proof was provided with the
discovery of the tombstone of T. Flavius Proculus, pr(inceps) praeJ(ectus)
134
Scord(iscorum)who had been grantedcitizenship by a Flavian emperor.80 This
tombstonealso revealedthat, unlike the nativepraepositiin Dalmatia,the heads
of the civitates were known as praefecti like their military forerunners.Addi-
tional confirmation that local self-governmentwas grantedunder the Flavians
comesfrom anothercivitas. In 195 I a tombstoneset up to his wife by M. Coc-
ceius CaupianusPR.C.B. was discoveredin the enormousvilla of Parndorfin the
Leitha district.81 From his name it follows that he had acquired citizenship
under Nerva. The abbreviation can stand either for pr(aefectus)or pr(inceps)
c(ivitatis) B(oiorum), and either expansioncould be defendedwere it not for the
fact that in the civitas Boiorum a surprisingly large number of Flavii had
acquiredcitizenshipbeforeNerva'sreign. They retainedtheir traditional Celtic
names (e.g. T. Fl. Cobromarus,T. Fl. Biturix, T. Fl. Samio, etc.) and were
apparentlyvery rich, to judge from the unusuallylarge numberof their slaves
andfreedmenreferredto on inscriptions.82 Sinceit is clear that the Boian aristo-
crats had beengrantedcitizenship under the Flavians it is not very likely that
one of their most respectedmembersdid not acquireit until the time of Nerva.
The interpretationof PR asprincepsis thereforepreferable.From this wholesale
grantingof citizenshipto the inhabitantsof the civitas Boiorum it follows that it
must have becomeself-governingat the sametime. In the early yearsof Vespa-
sian's reign the civitas was still under military control which at that time was
exercisedby VolcaciusPrimusin his capacityaspraefectusalae I Noricorum.83 His
headquarterswere in the ala's permanentfort at Arrabonawhich the Boian and
Azalian civitates adjoined. V olcacius was the praefectusof both civitates,which
perhapsindicatesthat military control was being relaxed.It was, however,still
in force under Vespasian.
It is possiblethat the abolition of military control over the civitates in south
and west Pannoniawas connectedin someway with the founding of towns by
the Flavians(Fig. 37, p. 220). ColoniaSiscia wasplacedin the territory of the civi-
tas Colapianorumwhich under Nero had hadAntonius Naso as praefectus;the
probable result of its foundation was that the territory of the civitas was
absorbedby that of the colony. The Flavian colonia Sirmium was foundedin the
territory of the civitas Amantinorum,andit is probablethat the territorium of the
Flavian municipium of Scarbantiawas carvedout of the civitas Boiorum, as was
the territorium of colonia Savariaat an earlier stage.Two southPannoniancivi-
tateswere raisedto the statusof municipia underthe Flavians,andwere the first
indicationsof a new policy of urbanizationwhich after Trajan completelysuper-
sededthe old systembasedon colonizationor settlersfrom outside.The munici-
pium Flavium Latobicorum (Drnovo) was founded close to the mouth of the
Korkoras (Krka); its name shows the connectionwith the local tribe of the
135
Latobici. Furtherexamplesof this type of namefor municipia in both Pannonia
and UpperMoesiaat a later datewill be quotedbelow (p. 223). It is probablethat
the new Flavian municipium was simply createdout of the civitas Latobicorum,
that the latter'sprincipes becamethe decurionsof the new town, and that the
raising of its status was coupled with a fairly wide grant of citizenship to its
inhabitants. Some Flavii are in fact mentioned on the not very numerous
inscriptionsfrom the vicinity.84 The namemunicipium Latobicorumwas soon
disused:on later inscriptionsit is called Neviodunumand as suchit is listed in
the Itineraries.85 There are many examplesfrom the middle Danubeareaof this
kind of place-namechange.Thelocal Celtic namefor the settlementnearDrnovo
was Neviodunumand continuedto be used,evenafter the areaof which Nevio-
dunumwas the centrehad beengiven a new official title. The latter was unable
to establishitself, a fate which was later sharedby such namesas municipium
Iasorumin Pannonia,colonia Ulpia Traiana Dacica (Sarmizegethusa),colonia
Flavia Felix Domitiana (Scupi), etc.
The other Flavian municipium in the upper Save valley was Andautonia,
which had beenfoundedin the territory of the civitas Varcianorum(SCitarjevo
not far from Zagreb)(Fig. 9). Although there is no evidencethat it was given a
new name,this is not unlikely. In any caseby the time of Trajan it was again
called Andautonia.86 Both Neviodunumand Andautoniaprobably owed their
relatively early emergenceto shipping on the Save.Although there is evidence
of citizenship being granted to the Celtic natives on the upper Save by the
Flavians and even earlier, and although the name municipium Latobicorumis
evidencefor the native origin of the settlement,a section of the urban upper
class was descendedfrom the Italian entrepreneurswho had settledalong the
Savein the courseof the first century. North Italian names(Aquileians) suchas
Marcii, Rustii, Capenii, Annaei, Trotedii, Firmidii, etc.,87 are epigraphically
attestedat Neviodunum,and from Andautoniathe Iuventii, Caesernii,Aconii,
Lusii, etc.88 Many of them were freedmen,or indeed slaves,as for examplea
certain .. .]medus Trotedi negotiatioris servus. It was significant that they set up
altars to the gods of shipping-toNeptuneas a generalrule, but also to Savus.
For instanceit was to the latter that a certainM. Iuventius Primigeniusmadea
dedicationalong with his businessassociates(socii). Inscriptionsset up by north
Italians and also by somesouthernGauls (Eppii) 89 comefrom the nearvicinity
of Neviodunumand Andautonia,whereasthoseof the nativeshavebeenfound
mainly on the peripheryof the territories.
Thus theseFlavian municipia were basedequally on the native population
and on foreigners. Neverthelessthe aim was clearly to createtowns from the
civitatesperegrinaeand, as will be seenlater, this aim both underand after Hadrian
136
becamethe guiding principle. The necessaryconditionsdid not, however,obtain
everywhere;thesewere obviously the existenceof a settlementof a more or less
urbannature,for which of coursetheforeign settlersprovidedthe main impetus;
the presenceof a local upperclasswhich not only numericallybut alsofinancially
andpolitically was capableof forming an urbanaristocracy;andfinally a certain
numberof Romancitizenson whom the early stagesof urbanlife would depend.
Most of these conditions were to be found only on the south and western
boundariesof Pannonia,the areasof the mostimportanttraffic routesin the early
period. That administrativemeasureswere necessaryevenherein order to open
up shippingon the Savehasalreadybeenmentioned(p. I 13) in connectionwith
the foundation of the coloniesof Siscia and Sirmium.
The conditionsjust mentioneddid not exist on the banksof the Danubeuntil
the late Flavian-Trajanicperiodandevenlater in the interior of the province.As
we have seen,large settlementswere establishedby foreign tradersnearlegion-
ary fortresseson the Danube,and we may suspectthe beginningsof a similar
processat various points along the diagonalroadsfrom Poetovioto Aquincum
andfrom Mursa to Savaria,for instanceat Sopianae,althoughfrom the outsetit
was hinderedand finally broughtto a standstillby the developmentof tradeon
the Danube. The beginnings of systematicgrants of citizenship both in the
interior of the provinceand on the Danubelikewise go back to the late Flavian-
Trajanicperiod.The Flaviansbestowedcitizenshipon the nativearistocracyonly
in certainareas.Evenin the municipiumLatobicorumthe majority of the natives
remained peregrini; they are mentioned on numerousinscriptions from the
westernhalf of the urban territory.90 It is only with the appearanceof Ulpii of
native origin that we cantracea universal,albeit gradual,grantingof citizenship.
TheseUlpii are mainly attestedin the following civitates and towns: Eravisci,
Azali, Boii, Andizetes,Scordisci,Iasi, Breuci, Cornacatesand Savaria,Sirmium
andPoetovio.Probablythe first to embarkon this stepwas Nerva, who granted
citizenship not only to the Boian princeps Caupianus,as already mentioned
(p. 135), but also to Florus, a princepsof the Eravisci and son of Matumarus.91
Often it is not possibleto distinguish betweencivilian membersof the upper
classwho hadbecomecitizensandveteranswho hadbeensimilarly honoured.In
north-easternPannoniathere is clear evidencethat citizenship was grantedto
civilians.92 In the long run it is of courseirrelevantwhetherthe daily increasing
numberof natives with Romancitizenshipwere civilians or veterans.Trajanic
new citizens (Ulpii) or their descendantsare attestedin all the areas where
Hadrianfoundednew towns.
The situationwas completelydifferent in UpperMoesia. Neither the Flavians
nor Trajanfoundedmunicipia there; moreoverthey showedconsiderablereluc-
137
tancein grantingcitizenship.It is not in fact surprisingthat municipia could not
be founded, since, as we have seen,settlementson a large scale only beganto
develop in the reign of Trajan, while native civitates had not yet reachedthe
stagewhenmunicipalizationcould takeplace.As alreadypointedout in Chapter
3 the civitatesperegrinaein Upper Moesia were reorganizedin the secondhalf of
the first century, doubtlessas the result of massivesettlementsof barbarians
from the left bank of the Danube;thesehad probablynot reachedyet the stage
at which any relaxationof strict military control would havebeenpossible.Nor
was it accidentalthat the mines were first developedunder Trajan; it obviously
was not as thoughthe presenceof mineralshad just beendiscovered.Even with
mines of suchgreatimportanceto the empire,securitywas the prime considera-
tion. Thereis no evidenceat all beforeTrajanfor a grant of citizenshipin Upper
Moesia. Even he grantedit only in a single restrictedarea,that of the Hadrianic
municipium of Ulpianum (Fig. 25, p. 146), coinciding roughly with the autono-
mous Jugoslaviandistrict of Kosovo and Metohija (Kosmet). This area, conti-
guouson the north with the Flaviancolony of Scupi,wasprobablywell pacified,
not only becauseit was neara colony but also becauseit was an extensiveagri-
cultural district, as the Greeksof the classicalperiod knew.93 It is preciselyfrom
this areathat one of the few latifundia (huge estates)on the Danubeor in the
Balkans is attested;it belongedto the senatorialfamilies of the Furii and the
Pontii, the latter having acquired it at the latest under Hadrian.9 4 A further
factor was that the road from the Adriatic to Dacia passedthroughthe Kosovo
Polje. The Trajanic new citizens in the Kosmet, unlike the Boian Flavii, mostly
had nondescriptLatin cognomensbut the namesUlpius Andinus, Ulpia Andia
and M. Ulpius Timentis f. Maximus, as well as the single Illyrian or Thracian
namespossessed by relativesof the Ulpii, indicatetheir local origin.95 Thereis,
incidentally, no information concerning the fate of the civitates peregrinae in
Upper Moesia.96 It is, however, probablethat the civitas Dardanorum,whose
territory roughly correspondedto the Kosmet, becameself-governing under
Trajan at the sametime as its inhabitantswere grantedcitizenship.
Hadrian'sprovincial policy on the Danube,it would seem,was to continue
and concludethe developmentstartedby his predecessors, particularly Trajan.
Even if we confine our attentionto the numberof municipia Aelia in Pannonia,
the importance of Hadrian in the history of the province clearly emerges.
According to presentevidencehe createdeight municipia (Fig. 37, p. 220),
thoughit shouldbe emphasizedthat this may well not be the final figure. Three
out of the sevenbecameknown only a few years ago, hencethe discovery of
further municipia createdby Hadrianor evenby later emperorscannotbe ruled
out. He also createdat least two municipia in Upper Moesia.
138
It is quite possiblethat the municipia in Pannoniaand Upper Moesia were
founded during Hadrian'svisit there in 124. He reachedthe Danubefrom the
east via Thrace, and from Pannoniacontinuedhis journey in the direction of
Dalmatia. A few minor incidents which occurred during his progressare re-
cordedon inscriptions;for instance,accordingto an epitaphto a favourite horse,
possibly composedby himself, he took part in a boar hunt in the wilds of
Pannonia.97 He also had a trooperbelongingto the cohorsBatavorumreport on
his outstandingdeedsof bravery.98 An inscriptionin honourof Hadrian,incised
on what was probablythe plinth of a statue,suggeststhat Aquincum was raised
to the statusof a municipium during an imperial visit. 99 The other municipia on
the Danube,Viminacium and Carnuntum,wereprobablycreatedin like manner.
Thesewere, of course,the seatsof the governorswhom Hadrianhad doubtless
visited.
Canabaeand Hadrianic municipia

The three municipia createdby Hadrian on the Danube (Carnuntum,Aquin-
cum and Viminacium) put the final touches to urban developmentnear the
legionary fortressesbegun under the Flavians or even earlier. By the time of
Trajan, at the latest,this developmenthad producedthe corporateautonomyof
the cives Romaniconsistentes.While the topographyof the 'camp towns' on the
Rhine and Danubestill remainedobscureit usedto be thought that the muni-
cipia foundedin the vicinity of the fortressescould be identified with the settle-
ments there which had developedinto towns with the cives Romani or the
canabaeas their pre-municipalform of administration.But after it becameclear
that nearalmosteverylegionaryfortresson the Rhine andDanubetherewas not
one but two civil settlements,the problem becamemuch more complex and
many of its aspectsstill await clarification. The following topographicaldetails
aretypical of Danubelegionarybases(Fig. 22, p. 127): aroundeachwasa civilian
settlement,the creationand developmentof which ran parallel with thoseof the
fortress itself. The Roman citizens who had establishedthemselvesin this
settlementformed the pseudo-autonomous corporatebody of cives Romanicon-
sistentesad legionem. This civil settlement,later called canabae,sharedthe neigh-
bouring cemeterieswith the military. At least one Romanmile from this settle-
ment and from the fortress there was another civil settlementwith its own
cemetery.Betweenthe two settlementswas a spacewhich wasneitherbuilt upon
nor usedas a burial ground.It was this secondsettlementat a distancefrom both
canabaeand fortress which becamea municipium, whereasthat near the camp
continuedto have a semi-municipalautonomywhich in legal terms was quite
139
different from that of the municipium. That the two civil settlementswere on
different sites has beenclearly establishedat all the fortressesin Pannoniaand
Upper Moesia. At Carnuntum,Brigetio, Aquincum and Viminacium the settle-
mentswhich sooneror later becamemunicipia were abovethe fortress,either to
the west or north, whereasat Vindobonaand Singidunumthey were below it,
that is to the east.The reasonsfor thesedoublesettlementswould not be difficult
to identify if thosewhich becamemunicipia under Hadrianhad also beenfoun-
ded underhim. In that casethe duality could be explainedby the fact that settle-
mentscloseto fortressescould not be municipalizedsincethey occupiedground
belongingto the military, where no municipal territorium could be provided.
Hadrianwould then have had a town foundedat somedistancefrom the camp
and have granteda municipal charter;in so doing perhapshe might also have
wishedto checkthe developmentof the settlementadlegionem.The matteris not,
however,as simple as that. Investigationshave alreadyproved that the settle-
mentswhich becamemunicipia underHadrianor later were alreadyin existence
at the time of their promotion and were probably founded at the sametime as
the settlementsad legionem,or not very much later.lOo
Our starting-pointmust be the territorial situation.It is usually assumedthat
settlementsad legionemwere established-justas the fortressesthemselves-on
land which was not only owned but also administeredby the military. Prata
legionis (pasture-landsbelonging to the legion) are in fact attestedin the first
centuryin Spain,Dalmatiaand possibly also in Lower Germany.On this land,
where territorial sovereigntywas in the handsof the military, municipal auto-
nomy was impossible,and it follows that Romancitizens who had settledthere
could only form a corporatecommunity, that of cives Romaniconsistentes.Unfor-
tunately there is no information concerningthe size of the prata legionis on the
Danube.Thereis, however,a strongpossibility that the settlementswhich later
becomemunicipia did not lie on military ground. The fact that two civil settle-
mentswere establishednearthe campsmay bestbe explainedby the supposition
that the circumstancesunder which the settlersestablishedthemselvesand the
communallife in the two typesof settlementdiffered right from the start; more-
over, the entrepreneurs,soldiers'relatives,veteransand other foreignerscould
chooseto settle where the legal and administrative conditions seemedmore
appropriateto the kind of life they had in mind. When Hadrianbroughturban
developmentnearthe campwithin the scopeof public law he could only raisethe
settlementwhich lay at somedistancefrom the campto the statusof a municip-
ium, for otherwisehe would have had to deprive the military of control over
the area closest to the fortress. Septimius Severuswas the first to take this
step.IOI
140
Thusthosesettlementswhich becamemunicipiaunderHadrianwereprobably
on territory belongingto the neighbouringcivitasperegrina; andthereis much to
supporttheview thatthelatterwasmunicipalizedat the sametime. This emerges,
for example,from the fact that the decurionsof the municipiumAelium Carnun-
tum have left inscriptions in the former territory of the civitas Boiorum, and
some of them had acquired citizenship from the Flavians or Trajan.102 It is
probable, nevertheless,that the ordo of Carnuntumwas largely composedof
foreigners who had beenliving for a long time in the vicinity of the fortress.
New settlersare also attestedin the ordo: the decurionC. Domitius Zmaragdus,
who was sufficiently wealthy to pay for the building of the amphitheatreat
Carnuntumout of his own pocket,was an immigrant from Antioch and one of
the earliest Syrian tradersto settle in Pannonia.103
The ordo of municipium Viminacium likewise consistedof foreigners,exclu-
sively so, it would seem.I04 It is from here that there is definite evidenceof a
decurionbeing the son of a freedman.lOS Freedmenwere doubtlessone of the
most important componentsof the consistentesad legioJ1em in the Flavian or
Trajanicperiods.In towns wheretherewas a flourishing economiclife it was the
freedmenwho had grown rich who later provided the urban aristocracy.
The developmentat Aquincum was entirely different. Doubtless here too
Hadrianintendedto build up the life of the new municipium on the basisof the
wealthiestinhabitantsof the settlementad legionemor, more precisely,to transfer
their not inconsiderablewealth to the new municipium; but in this, it would
seem,he was unsuccessful.The by no meanssmall numberof decurionsof the
municipium known to us from inscriptions in the period from Hadrian to
Marcuswere without exceptionnativeswho owed their citizenshipto Trajan or
to his successors,Hadrian and Pius (Fig. 25).106 Foreignersdid, of course,play
their part in the life of the town as membersof the numerouscollegia,ro7as did
the veteransof the legio II Adiutrix, but they were either not representedin the
ordo at all, or if so only to a limited extent. Despite everything, the civitas
Eraviscorumwas not dissolvednor did it disappearwithout trace in the newly
founded municipium Aelium Aquincum: it survived after Hadrian'stime and
was even mentionedon an altar of the third century.lOS
The rich epigraphicmaterialfrom Aquincumandits environsprovidescertain
clues enablingus to reconstructa situationwhich was legally extremelycompli-
cated.In the first centurythe civitas Eraviscorumhadprincipeswho were called
decurionesunder Trajan.ro9 Under Hadrian and later, however,there is evidence
of only one tabulariusIIo and one functionary armCo ..),III the latter otherwise
attestednowhere else; hence some change in the status of the civitas must
definitely havetakenplaceunderHadrian.In the sameareawheretheprincipesor
141
decuriones,tabularii and the other ordinary membersof the civitas Eraviscorum
are epigraphicallyattested,inscriptionsto the decurionsof the municipiumhave
also beenfound. Consequentlythe latter ownedpropertyin an areawhich up to
the time of Hadrian had belongedto the civitas Eraviscorum.Neverthelessit
cannotbe assumedthat the territory belongingto the civitas was mergedin that
of the municipium, becausethe communityitself, that is the civitas, althoughit
had undergonea change,persistedeven after Hadrian. Since the decurionsin
Aquincum camefrom the territory of the civitas, it must be assumedthat the
latter was in some sort of position of dependenceon the municipium. The
probability is that the administrationof the civitas was taken over by the ordo,
the membersof which were, of course,well equippedto representthe civitas,
since some of them came from it. The civitas which continuedto exist as a
territorial entity had only a few administrativeofficials (tabularii), and perhaps
one or two elected honorary heads (arm . .. ?) who were at the same time
decurions.The inhabitantsof the civitas, not unlike those in Latin municipia,
wereableto acquirecitizenshipon becomingmembersof the ordo, or as auxiliary
veteranssincerecruitmentwas continued:diplomasmadeout to Eravisciin the
post-Hadrianicperiod are not infrequent.II2 Basically, the relationshipbetween
the municipium and the civitas resembledthe attributio, thoughthis term is not
attested.Thus it was naturalfor two duoviri from Aquincum, both nativeswho
had recently acquired citizenship, to set up an altar to the well-being of the
civitas Eraviscorum,II3 and since the latter was never formally abolished it
could be commemoratedby an auguron an altar in the third century.II 4 At that
time it probably existed only as a community basedon a religious cult, since
betweenthe time of Hadrian and that of Caracallaall legal and administrative
differences betweenthe two kinds of community had probably ceasedto be
significant.
It is possiblethat Hadrianalso reformedthe administrationof the settlements
ad legionem, as it was under him that a changein their corporateorganization
actually took place. On inscriptions of the second century the settlementat
Aquincumis calledcanabae;I I 5 althoughthis term was not appliedeverywhereas
a regular thing to thesesettlementsin the secondcentury, it was more often
used than the old term cives Romaniconsistentes.The latter had had curatores or
magistri, but the canabaealwayshad decurionsin addition and all the magistrates
found in a municipium or colony. This pseudo-municipalform symbolizedthe
urban characterof the 'hutted settlement'(canabae) but at the same time it
representeda compromisemade necessaryby the fact that the territory on
which the canabaewas sited belongedto the military. Thus during the second
centurynearthe fortresseson the Danubethereweretwo separateself-governing
14 2
The first age ~f prosperity
communitiesand actually threeat Aquincum, wherethe municipium,the civitas
Eraviscorum and the military each had their own territories. That of the
municipium was probablyto the west and north of the town, that of the civitas
peregrina to the south and south-west.Where the military territory lay, lack of
information makesit impossibleto say. It may, however, be assumedthat the
land round the auxiliary forts was part of it. An inscription records the cives
Romaniconsistentesnear the fort of Vetus Salina.II6
Hadrian'sothermunicipiawere sitedin the interior of the provinceand hence
hadtheir origin in purely civilian developments.Two typescanbe distinguished
-thatwhich developedout of a civitasperegrinawhosedecurions,it would seem,
were natives who had recently acquired citizenship; and that which granted
autonomyto groups who were alreadycitizens, and in part at least of foreign
origin. In making this distinction it would, of course,be wrong to insist on
consistency;at the momentit is still not possibleto say whetherin the first type
of municipium foreignersalso had a say in affairs.
Includedin the first type in Pannoniawere municipium Iasorum,Cibalaeand
Bassiana,and in Upper Moesiamunicipium Ulpianum. Thosein Pannoniawere
on the road from Siscia via Mursa to Singidunum,and as the result of their
foundation the whole of the Sirmian region and the greaterpart of the Drave
and Savevalleys were municipalized.Municipium Iasorumwas situatedroughly
midway betweenSiscia and Mursa at the spa-siteof AquaeBalizae(Daruvar).II7
The official name given to it is a clear indication that the civitas Iasorumhad
becomea municipium. Evidencethat its territory was extensiveis providedby
an inscriptiongiving the origo of an equessingularisat Rome,who hailedfrom the
vicus Iovista on the Drave, which was said to belong to Aquae Balizae, other-
wise called municipium Iasorum.II8 The name municipium Iasorum is also
attestedon a tomb-altarof a decurion,but as with the namemunicipium Lato-
bicorum it failed to catchon. In the AntonineItinerary the placeis called Aquae
Balissae.Cibalaelay southof Mursa (modernVinkovce) at a naturalintersection
of routesandin an areain which practicallyall the importantbattlesin Pannonia
took place.II9 The areabelongedto the civitas Cornacatium,the former centre
of which, Cornacum(Sotin), was not far from Cibalaeon the Danube.Bassiana
(Petrovci), between Sirmium and Singidunum, was in the territory of the
civitas Scordiscorum.I2o
Only one decurionfrom eachof thesethree Hadrianic municipia is attested,
namely P. Aelius P. fil. Aelianusscriba, decurio and IIII vir of municipium Iaso-
rum,IZI M. Ulpius Fronto Aemilianus, decurion of CibalaeI22 and P. Aelius
Dasius,decurionof Bassiana.I23Despitethe sparsedetailsit is hardly accidental
that the grantingof citizenshipto thesedecurionsdatesto Trajanor in two cases
143
to Hadrian. The ordo of thesemunicipia was obviously formed from the aristo-
cracy of the civitas peregrina. Mention should also be madeat this point of the
municipium Mogentiana, since it also had Aelii among its decurions.124 Its
territory was, however, north of Lake Pelso, where there is evidence that
foreignershad settledat an early dateand ownedlarge estates.Thus a decurion
was called Cassiusand a scriba possibly Dubius.I2S Unfortunatelyit has so far
proved impossibleto locate Mogentiana,thoughit lay somewhereon the road
between Savaria and Aquincum. Taking into account the contradictory dis-
tancesgivenin the AntonineItinerary, its site shouldbe west or north-westof the
Bakony mountains,the most likely places being Somly6vasarhelyor Sumeg-
csehi. A further difficulty in locating its site is that, apart from one tombstone
with a doubtful reading(scriba municipii M.), all the epigraphicevidencecomes
from areas which definitely did not belong to Mogentiana'sterritory, as for
examplethe neighbourhoodof Brigetio, Aquincum and Tokod on the Danube,
eastof Brigetio.126
The municipiumMursella, too, which was probablyalsofoundedby Hadrian,
was situatedin an areain which foreignersand veteranshad settledin the first
century.Its territory probably took in both banksof the Arabo (Raba)southof
Arrabona(Gyor). Thereis epigraphicevidenceof two of its officials, a scriba by
the name of Gallonius and a decurion called Iulius who belonged to the
EquestrianOrder.I27
Recently a new inscription came to light at Aquae Iasae,128attesting a
Hadrianic municipium namedSalla. Salla (or Sala) is probably to be found at
Zalalovo or at any rate somewherein the valley of the river Zala, where the
place-nameSalla is recordedon the Poetovio-Savariaroad. The territory of
this small insignificant city lay betweenthose of Poetovioand Savaria.
A striking fact is that in the towns of the interior of Pannoniaalmostas many
scribae are attestedas are decurions;in the municipium Iasorum there was a
scriba who wasalsoa decurionandevenbecamea memberof the Quattuorvirate.
As a rule the scribae were paid employeesand thus were not membersof the
wealthy urban aristocracy. Where in these and other Pannoniantowns (for
instance municipium Faustinianum)they appear to possesswealth and also
belonged to the aristocracy, there must have been cogent reasonsfor their
exceptionalstatus.We have seenthat Hadrian'spolicy of urbanizationaimedat
involving the local nativearistocracyin the administrationof the new municipia:
in more generaltermsthe nativecommunitywas to be directedalong the pathof
municipal development.It is possiblethat the aristocracywas not equal to this
task, and in consequencethe scribae who held the chief posts within the urban
administration acquired added importance. They may well have headedthe
144
administrationand consequentlypossesseda much higher social position than
they would haveenjoyedin largertowns,wherethey would haveremainedpetty
officials in an administrationcontrolled by the aristocracy,129It is also worth
noting that the inscriptions of decurionsfrom Salla, Mursella and Mogentiana
were found neither in the towns, nor even in their territories. This paradoxical
situation is possibly attributable to the fact that the decurions were hardly
involved in the public life of the town. If we rememberthat in both Mursella
and Mogentianaland and wealth were firmly in the hands of foreigners even
before they became municipia, it becomesunderstandableif the decurions
showedlittle interest in participating in the life of the town, the founding of
which had brought them no real advantage.
The situationwas also similar in the only municipium foundedby Hadrianin
the interior of Moesia; Ulpianum near Gracanica,not far from Pristina.130 It
must have been connectedin some way with the metalla Ulpiana, otherwiseit
would not havebeencalledafter the mines.Lessees,minersandthe mine admini-
stratorsprobably formed the nucleusof urban life, since the town was not far
from the mines on Mount Zegovac.Nevertheless,the decurionsso far attested
were Ulpii and Aelii, whose inscriptions were significantly found not in the
town but west of it in the Metohija,131 where there is considerableepigraphic
evidenceof citizenship-grantsto natives, both under Trajan and later. Mean-
while a peculiarity, also characteristicof Aquincum, shouldbe noted: the Ulpii
are found mostly in the territory, whereasthe Aelii tendedto live in the town
(Fig. 25),132 Thus Hadrian'spolicy of granting citizenshipwas harnessedto the
serviceof urban development.
Even after his intensive policy of municipalization some civitatesperegrinae
remaineduntouchedby it. In PannoniaSuperior the Cotini (K ytnoi), Azali,
Serapilli and Oseriatesand in Inferior the Hercuniatesand Breuci were still
civitatesperegrinae;at any rate thereis so far no evidencethat they beomemuni-
cipia under Hadrian. In Moesia Superiorprobably all the civitatesperegrinaeper-
sistedas such after his day, with the possibleexceptionof the Dardani, whose
area largely coincided with the territory of Ulpianum and Scupi. The two
coloniesand the two municipia in Moesia Superioroccupiedless than half the
province. Viminacium's territory was remarkablysmall.
Pannonia'snative populationleft behind a great variety of late first-century
and second-centuryarchaeologicalmaterial, which deservesspecial attention,
principally becausethereare markeddifferencesbetweencertainzoneswhich do
not fit into the processof municipalizationdescribedabove.Even after making
allowancesfor the tardinessof investigationin somesouthPannonianareas,it is
neverthelessa striking fact that no wealth of epigraphicmaterial,or evidenceof
145
OVLPIANVM
o, so
km
VLPII
AELII
++
++
o SO
• I
+
km
OVLPIANVM
++++
+
+++ •
++
Figure 25 Ulpii and Aelii in the territorium of Aquincum and Ulpianum
146
certainspecialburial customsindicative of a prosperousor wealthy native class,
has normally beenfound in the territories of the municipia foundedunderthe
Flaviansand Hadrian;on the contrary,amongthe nativesRomancustomsseem
to have beenmore deeply rooted in areaswhich were not municipalizeduntil
later in the secondcentury,or evensubsequently.Evidenceof this sort is scarce
in Flavian municipia. Before this problem can be dealt with in any detail it is
necessaryto consider an apparent paradox. Local (non-Latin) names, local
costumeand sometraditional burial customsare to be found in the very areas
where Romaninfluence on the natives was strongest.Attention has frequently
been drawn to these local peculiarities simply as proof of the weaknessand
superficialnatureof Romaninfluence.The stubbornpersistenceof nativenames,
for example,it was urged, was the result of this influence being insufficiently
powerful. The right view is, however, precisely the opposite. Celtic, I1lyrian
(Pannonian)or Thracian names have come down to us becausethey were
recordedon Latin epitaphs.If the nativeshad not takenover the Romancustom
of settingup tombstones,the identification of traditional or Latin nameswould
have beenimpossible.In the sameway the fact that the natives carvedreliefs
of their dead on tombstones providesclear evidenceof persistingfashion in
women'sdress.Thus native dress,native namesand many other aspectsof local
culture could only be transmittedvia Romanforms; traditional elementscould
be expressedonly throughthe interventionof externalmedia.It canprobablybe
assumedthat Romanizationof the local mode of life and way of thinking was
leastevidentwheretherewere neitherRomantombstonesnor otheradjunctsof
Romanculture.
Burial rites, tombstonesand their social significance

The samealso holds good for some native burial-rites and their ideas of life
after death. The so-called astral symbols (Fig. 10), already touched upon in
Chapter3, were probablyof ancientorigin, but are first met with as embellish-
ments on tombstones.The most impressiveexample,however, is that of the
cart-burials. The journey of the dead to the next world in carts, boats or on
horsebackis an ancientand widespreadconcept.Vesselsshapedlike carts,parts
of vehicles found in graves and isolated wagon-burialsare evidencefor the
currencyof theseideasin both the earlier and later periodsof the Iron Age.133
The sameconceptalso found expressionon tombstonesin Roman times; the
dead man is depictedin relief in a four-wheeledvehicle drawn by a team of
horsesor oxen.I34 Suchtombstonesmadetheir appearance beforethe end of the
first century,were fairly numerousin the secondand only ceasedat the time of
147
the Severi. Most of the dead, particularly those commemoratedon the earliest
tombstones,were natives,mainly peregriniwith single, Celtic names.In the first
half of the secondcentury the customof wagon-burialsbecamefashionable.I3S
The deadman, depictedin relief on the tombstoneas travelling in a vehicle, was
buried togetherwith a four-wheeledwagon, horses,a set of sacrificial utensils
(tripod, pitcher and plate), and with certain luxury objects such as equipment
from a palaestra(aryballosand strigil), iron folding chairs, etc. (PI. 15). Every-
thing speaksof an extraordinarydegreeof luxury; the vehicle was decorated
with bronzereliefs and had rich mountings,the horse'sharnesswas definitely of
a ceremonialcharacterand the rest of the equipmentwas all imported, mainly
from the bronze workshopsof the Rhineland, but also from the easternpro-
vinces. Pictorial representationswere mostly taken from the Dionysiac cycle.
That the richestnativeswere buried in thesegravesis substantiatednot only by
the fact that there is evidencethat the idea of the journey to the next world by
wagon pre-datesthe actual wagon-burials(scenesdepicting wagons on tomb-
stonesand pre-Romanfinds) but also by the distribution of thesewagon-burials
in preciselythoseareas(including thosewhereforeignersplayeda leading role)
where tombstoneswith wagon-reliefshave not come to light.
There is no doubt that these wagon-gravesoriginated from a very wealthy
native aristocracystill very much tied to tradition, which, althoughits way of
life as far as externals were concernedhad been largely Romanized,had not
abandonedits ancient beliefs about the next world. In fact it may be said that
thesewagon-burialsonly becamepossibleas the result of Romancontrol, which
led to the aristocracyacquiringwealth and adopting a civilized 'Roman'way of
life. The man buried in the wagon-graveat Sarszentmikloswas probablyone of
the highest functionariesof the civitas Eraviscorum,since his grave also con-
tained a lance-shapedbadge (hasta) made of bronze, to which only represent-
ativesof executivepowerwereentitled.136 The ancientview that the deadentered
the next world in vehicles was expressedin visual terms from the moment the
nativesbeganto set up tombstones,but to put 1t into practice by actually dis-
patchingthe deadin a wagon was possibleonly for the wealthiest.
Although this form of burial was probably restrictedto the tribal aristocracy
who owned vehicles, burial under tumuli was a socially less restrictedcustom.
The indication of a grave by a tumulus is such an obvious idea that it was
adoptedeverywhereand at all times. Tumuli were customaryin Pannoniain the
early part of the Iron Age, occurred sporadically in its later stagesbut were
mostfrequently employedin the Romanperiod. Thesetumuli, some1-3 m high
and 5-12 m wide, were raised over graves containing crematedremains, and
beganto appearaboutthe middle of the first century in westernPannoniaand
148
km· km· km· km·
Native on
inscription
Wagon- grave
Chariot-scene
on tombstone
Tumuli
km·km·
km·
km·
Figure 26 Monumentsto natives in Pannonia
aboutthe beginningof the secondcenturyin the easternpart of the province.I 37

This too suggeststhat once conditions had been consolidatedold customs
flourished anewand evenbecamemore widespreadthan before. Of coursethis
appliedonly to areaswhere the native populationhad reacheda certain level of
prosperity. Hence the ability to establishwhere thesecustomsprevailedis not
unimportant.
Attention has been drawn to some specialfeaturesof native culture for the
purpose of showing that they derive from traditional customs and ideas of
149
which we shouldhaveremainedignorantbut for the effect of Romancontrol. A
distinction must, however, be made. There were somelocal, non-Romancul-
tural aspectswhich were widespreadbut which requiredRomanforms for their
expression.That native (Celtic or Illyrian) nameswere not abandonedin the
first and secondcenturiesfollows a fortiori from the fact that they appearon
Latin epitaphs.Similarly, the ideaof conveyanceinto the next world in a vehicle
was probablywidespread,but we are only madeawareof it from portrayalson
tombstonesor in someotherpracticalmanner.The style of nativewomen'sdress
obviously persistedeverywhere,but again our only information concerningit
comesfrom reliefs on tombstones.In all theseinstancesthe habitof settingup
tombstoneswas a symptomthat Romancustomshad beenadopted,and at the
sametime it provides information concerningwidespreadnative customs.On
the otherhandtherewere aspectsof this culturewhich owedtheir dissemination
to the Romans;wagon-gravesand tumuli, while they reflect ancientattitudesof
mind, were first given practicalexpressionwhen the nativescameunderRoman
control, and then only where there was no social handicap.
Thesenative peculiaritiesof provincial Romanculture which had their roots
in native tradition are often regardedas symptomsof the so-calledCeltic, or
Illyrian, Renaissance.From what has beensaid, however,it will be seennot as
the result of any decline in or relative weaknessof Roman culture, but rather
that Romaninfluence either gave expressionto or revived local characteristics.
It is highly significantthat in Pannoniaall theselocal characteristicshavebeen
found either in associationor in the samearea. Henceit may be assumedthat in
theseareasthe native populationwas either under no social handicapor, what
comesto the samething, was ableto adaptits intereststo thoseof its rulers. The
demandswhich Roman provincial policy made on the ruling class were fre-
quently a clear indication of the latter's wealth and loyalty which found ex-
pressionin their singularcustoms.Epitaphspresupposea certainwealth on the
part of the families, but they also indicate the tendencyto copy the Roman
customof setting up tombstones.Wagon-gravesare not only a sign of wealth
but also reveal that certain externals of Roman culture had been taken over
(palaestraequipment,sacrificial utensils, etc.), though at the sametime many
traditional featuresof the native way of life and ideas were not abandoned.
It cannot, of course, be denied that certain differencesexisted in the areas
where this loyal native upper class is attested(Fig. 26). Tombstonesdepicting
the deadbeing usheredinto the next world in wagonsand actualwagon-burials
are almost evenly distributed. In the Leitha district, however, only reliefs of
wagonshavebeendiscovered,but no wagon-graves.The Boian aristocracywas
also familiar with the notion of the deadjourneyinginto the next world, but it
15 0
was not preparedto give practicalexpressionto it in the form of the expensive
and grossly extravagantcustom of wagon-burial, probably becauseliving as
they did in the hinterland of Carnuntumthey had long been in contact with
Roman culture. Even in the first half of the first century they had become
accustomedto Romanimportedgoods,they kept importedslaves,built villas at
leastby theendof the first century,were grantedcitizenshipandself-government
underthe Flaviansandwere ultimately mademembersof the ordo in Carnuntum
underHadrian.Thus they could look back on a long 'Roman'past,whereasthe
Eravisci, who did not becomeinvolved in Romantrade or play any important
part in provincial politics until the endof the first century,probablyclung more
persistentlyto their traditions. The majority of wagon-burialscome precisely
from this civitas of the Eravisci. To the southof it in the civitas Hercuniatium,
though wagon-gravesoccur, there are no tombstonesto natives. The richest
membersof the tribe were probably culturally and materially in a position to
allow themselvesthe luxury of theseburials, but the upper class in the main
lackedeither the necessarywealth or any cultural or political inducementto set
up tombstoneson Romanlines.
The areasin which, on the basisof what hasbeendescribed,the native aristo-
cracy was on the point of becominga Romanprovincial aristocracywere oddly
enoughnot so much the territories of the Flavian and Hadrianic municipia but
certainzones,the boundariesof which did not always coincidewith thoseof the
self-governingcommunities(civitates and municipia).
A smallishgroup of tombstonescomesfrom the native Celtic Latobici in the
westernpart of the territory of municipium Latobicorum.I38 Somewhatto the
north of that beginsthe broadzoneof the westPannoniantumuli, which includes
the whole of the Norican-Pannonianfrontier area from the Drave to the
Danube(Fig. 26). The only parts where there are no tumuli are where there is
epigraphicevidencethat foreignershad settledin the first century,namelyclose
to Poetovio(colonia Ulpia), Savaria(colonia Claudia)and within the areaof the
Flavian municipium Scarbantia. The tumuli are not indications of humble
burials, since they often reveal a wealth of finds: for examplewalled chambers
are not unusual, some of them with a dromos in the Mediterraneanmanner.
Somechambershave inscriptionsrecordingperegrineCelts or even Celts with
Romancitizenship.I39Although tombstonesof natives occur only sporadically
in the tumulusareatheir significancelies in the fact that sculpturallythey do not
differ from thosefound in the towns. A beautiful exampleis that of the tomb-
stone of Quartus Adnamatifilius, which was found not far from Savaria and
depictsthe she-wolfand twins.I40 Tumuli are lessfrequentin the Leitha district;
in contrastthere are many more tombstonesof membersof the Boian aristo-
151
cracy, and thesefrom the point of view of their reliefs form a class by them-
selves.The over-all compositionwas, of course, borrowed from first-century
tombstonereliefs at Carnuntum,but the deadwere depictedin local dress.The
stonemasonsseemto have beenlocal men.
It was no accidentthat the whole of this west Pannonianzonecoincidedwith
the Amber Road area, where in the first century brisk traffic had developed
betweenItaly and the barbariannorth via Carnuntum.In this the natives were
involved: a negotiator namedAtta Bataionis filius is recordedon a tombstone
(PI. 5b), found north of Savaria,and the stoneitself is a moderatelysuccessful
productionof a Savarianworkshop.HI
The other zone in which natives set up tombstonesalso coincides with a
tumulusareaand lies within the distribution of wagon-graves.In the north this
zone beginsat the Azalian-Eraviscanboundaryareaand includesthe whole of
north-eastPannoniafrom the Danubebend to roughly the Mecsekmountains.
Therewere no tumuli, wagon-gravesor tombstonestherebeforethe last yearsof
the first century and they only becomenumerousin the first half of the second
century. Thus archaeologicalfinds suggestthat comparedwith west Pannonia
therewas a time-lag of somefifty years,a result which agreeswith the historical
developmentin this area. Progressdid not begin until the late Flavian period,
when trade on the Danube and along the diagonal road from Poetovio to
Aquincum startedto flourish and when two new legionaryfortressesand their
numerousauxiliary units were establishedalong the river.
Trade and the military, however, made up only one component of the
Romanizationprocess.As already seenin Chapter3, the Celtic and Celticized
natives of Pannoniawere better at exploiting the advantagesof Romanrule, a
circumstanceprobablyconnectedwith the social structureof the Celtic tribes. It
was no accidentthat the two zones in which the native population had been
Romanizedwere Celtic, nor, probably,that in the Azalian area,where,it will be
remembered,south Pannonianshad been settledby Tiberius, only the narrow
easternstrip closestto the Eravisci sharedin this process.
Everywhereelse the relevantarchaeologicalfinds which might testify to the
Romanizationof the native population are either extremely sporadicor com-
pletely absent.It is not surprisingthat in municipia whereforeign settlerswere
presentbeforetheir founding (Andautonia,Mogentiana,Mursella)thesemenset
the tone and becamethe leadersin the communities,but in municipia wherethis
was not the casethe failure of natives to appearin either the epigraphicor the
archaeologicalrecord must be attributedto the backwardnessof native society.
Thesemunicipia were, of course,in the Illyrian-Pannonianarea (municipium
Iasorum,Cibalae,Bassiana),and,as we havealreadyseen,not eventhe members
152
of their ordo were completelyRomanized;the control of municipal affairs was
obviously in the hands of foreigners and the scriba who had been specially
appointedfor the purpose(p. 144). Thus efforts on the part of the Romansto
transformthe tribal aristocracyinto a municipal one met with no success.
In UpperMoesiait was only in the territory of the municipiumUlpianumthat
the native populationleft any important quantity of epigraphicmaterial, and it
is only here that there is evidencefor upper-classnative membersof the ordo
(Fig. 25, p. 146). In all other urbanand non-urbanterritoriesthereis absolutely
no evidencethat the native populationwas involved in the developmentof the
province. Traditional religious beliefs and customsare not recordedon monu-
ments, extremelyfew native Thracianor Illyrian namesare to be found in the
whole of the epigraphic material from the province, and native dress is not
depictedon tombstones.In generalit may be concludedeither that, apartfrom
the municipium Ulpianum, the native population resistedRomaninfluence or
that it was socially handicapped.
The reasonsfor this are obvious. As we have seen, Rome proceededvery
cautiously in occupying and constituting the province of Moesia. No war of
conquestprecededits founding and for a long time the completionand occupa-
tion of the Danube road was Rome's only achievementin and on behalf of
Moesia. The country was not of prime importancefor Balkan commerce;the
nativesliving in the high mountainsand narrow valleys were both inaccessible
and bellicose, and were far from having an appeal for entrepreneursand
colonists.Romenaturally often intervenedin local affairs by establishingsettle-
mentsor by reorganizingthe civitates, but the exploitation of the mines, obvi-
ously of great importanceto Rome, was only put in hand from the time of
Trajan. On the otherhandthereare reportsfrom aboutthe middle of the second
centuryof latronesDardaniae;I42 thus evenat this late dateit was not possibleto
pacify the mountainousparts of the province.
The backgroundof theseUpper Moesian robbersis explainedby the social
conditionsof the natives.As early as Greektimes the Dardanianswereknown as
wild mountain shepherdswho kept a large number of peasantsin bondage
(p. 27). With the gradual developmentof urban centres,the opening of the
mines and the appearanceof the few self-governingtowns, the peasantswere
ableto free themselvesfrom the shepherdswho thengraduallytook to banditry.
Under such circumstancesstableconditionscould not be establishedand com-
promisebetweennative and Romaninterestsbecameeven less possible.
153
Army recruitment
Even regular recruitment of Upper Moesians into auxiliary units did not
begin until the reign of Marcus Aurelius.143 Under Tiberius the Romansprob-
ably still recalledthat the Pannonian-Dalmatian rebellionwas sparkedoff by the
recruiting orderedby Valerius Messalla. Even when Upper Moesia becamea
provincelevies were not raised,and later conditionsremainedso unstablethat
the governmentdecidedagainsthaving Upper Moesiansin the army. The only
unit in which they served was the ala VespasianaDardanorumraised under
Vespasian,which did not, however, maintain its connection with Dardania
through subsequentrecruitment.Men from the colonies at Scupi and Ratiaria
servedin the legionsin the secondcentury; but, apartfrom thesetwo instances,
recruitmenttook placeonly rarely andas an exceptionalmeasure:for example,it
happenedoccasionallyat Ulpianum144 or with sons of soldiers serving with
auxiliary units145 in the province, though these did not recruit from Upper
Moesiannatives, except someDardanians.
The UpperMoesianarmy of the secondcenturywas composedof legionaries,
abouthalf of whom wereforeignersandthe remaindernatives,and of auxiliaries
who were almost exclusively foreigners. The legionariesrecruited in the pro-
vince itself were mainly citizens of Scupi or Ratiaria, but not infrequentlypere-
grine incolae of these colonies were also enlisted and in consequencebecame
citizens.146 The foreign legionaries came from neighbouring provinces-
mainly, of course,from the unarmedprovincesof the Balkans(Dalmatia,Mace-
doniaandThrace)andfrom Asia Minor. Among the auxiliariesmostly easterners
are attested.Thereis evidencethat a considerablemilitary classwas developing
in this period, for sonsof legionariesvery often joined the samelegion as their
fathers.
Recruitmentin Pannoniawas entirely different. In the Flavian-Trajanicperiod
manyItalians, Gauls,SpaniardsandNoricanswere still servingwith the legions;
the composition of the auxiliary units was even more varied, dependingon
wherethe unit camefrom or, not infrequently, the areafrom which it was later
reinforced.Syrians,Thracians,DalmatianIllyrians, nativesof the Alps, Britons,
Spaniardsand Germansmade up a not insignificant part of the auxiliary per-
sonnel. The majority of thesesoldiers went with their units to Pannonia;later
recruitmentfrom the provincesor tribes after which the units were namedwas
customaryonly for those with a special tactical function, as for instancethe
Syrians of ala I Ityraeorum147 or someThracianunits.
The date at which the transition to local recruitment took place cannot of
coursebe given, sinceobviouslyit was not the result of a single decree.Citizens
154
The first age oj prosperif
of the Pannoniancolonieswerealreadyservingin the legionsin Flaviantimes,I48
but they nevercompletelysucceededin oustingthe foreigners.It did, of course,
happenthat a legion or a vexillation from it was sentas an expeditionaryforce to
anotherprovince, from which it returnedwith a number of soldiers recruited
there.Thus orientalsjoined legio XV Apollinaris during the latter'slong stayin
the east (from about 62 to 71),149 while a substantialnumberof Macedonians,
men from Asia Minor and Syriansjoined legio II Adiutrix when it was engaged
in putting down the Jewish rebellion of Hadrian's reign.Iso Moreover, pro-
vinceson the Danubeandin the Balkanswhich hadno legionsof their own also
provided recruits. The majority, certainly from the time of Hadrian, gradually
came to be drawn from citizens of the Pannoniancolonies, though both
PannoniaSuperiorand Inferior producedrecruits only for their own legions.
By the time of ClaudiusPannoniannatives were alreadyserving in the auxi-
liary units stationedin the province.The personnelof the units transferredthere
in the Flavian-Trajanic period consistedlargely of foreigners, but once their
deploymentwas stabilized they were supplementedlocally. Such evidenceas
thereis of this is, however,extremelysparse. Auxiliariesin the secondcentury
left very few inscriptions,but as againstthis we possessa substantialnumberof
military diplomas (Fig. 27, PI. 22). The Pannonianswho receivedthem in the
second century were without exception Azali or Eravisci, which cannot be
regardedas a coincidencein view of the numberof diplomaswhich have come
to light from this period, particularly in Pannonia.ISI At this time both tribes
still had their civitasperegrina, andthesewere also recruiting areasfor auxiliaries.
What is not clear,however,is why the othercivitatesdid not providerecruitsfor
the auxiliary units or indeedfor the legions. So far no diplomas madeout for
Hercuniates,Breuci, Oseriates,Serapilli, etc., have been found. In the Julio-
Claudianperiod the Breuci had provideda very large numberof auxiliaries and
there is also some slight evidence of first-century recruitment among the
Hercuniates.IS2 Around the turn of the centurytherewerelevies of auxiliariesin
the Flavian municipia and actually in Siscia. T. Flavius Bonio from Andautonia
was a trooper with the ala Frontoniana,1s3and the cives Siscii et Varciani et
Latobici who were serving with the cohorts of Upper Pannoniaare probably
thoseattestedunderTrajanas being with a vexillatio in Syria.IS4 The possibility
cannot be dismissedthat after Trajan men from the municipia were being
recruited into the legions, though naturally this cannot be proved. Although
underHadriantherewere no fewer than elevenmunicipia in Pannonia,inscrip-
tions giving legionaries'birth placesall show that without exceptionthey were
citizensof the colonies;it is thereforenot surprisingthat, accordingto the large,
almostcompletelist of thoserecruitedfor legio VII Claudiain 169, a very large
155
A.D. 45-96
97-180
9 7 -1 8 0
97-180
97-180
o 100
, 200 km'
I I
A.D. 97-180
9 7 -1 8 0
97-180
97-180
Figure 27 Auxiliary diplomata from Pannonia

numberwere from Scupi and Ratiaria but nonefrom the Upper Moesianmuni-
cipia ofViminacium andUlpianum.I55 It hasalreadybeenpointedout that under
Trajan there was an equal number of colonies and legions in every province,
possiblybecauserecruitmentto the legions was restrictedto the colonies.What
is not clearis why the citizensof the municipia and of somecivitatesperegrinaedid
not serve in the auxiliary units either. Possibly the reasonlay simply in the
transitionto local recruitment.The auxiliary units in Pannoniawerestationedon
the limes sector,most of which lay on the edgeof Azalian and Eraviscanterri-
tory. The civitates of thesetwo tribes provided by far the largest number of
recruits as, of course, the other civitates whose territories extendedto the
Danubehad only one or two auxiliary forts in their vicinity. The long stretchof
the river below Aquincum, which was at the sametime the easternboundaryof
the Eravisci, perhaps explains why their civitas was not incorporatedinto
Hadrian'smunicipium of Aquincum but continuedto exist in an odd state of
dependenceon the latter. On the limes sector of the civitas Eraviscorumalone
there were no fewer than nine or ten auxiliary forts.I 56
What has beensaid in no way affects the fact that during the secondcentury
men from the canabae were recruited in increasingnumbersinto the legions.
Obviouslyonly sonsof soldierswere enlisted.Thesewere mostly both formally
andlegally citizensof the colony from which their parentshad comeor to which
they legallybelonged.I57 Althoughit cannotbe provedstatistically,thereis never-
thelessthe strongpossibility that the sons of these soldiers formed an increas-
ingly large part of the legions' personnel;by the end of the secondcenturyan
hereditary military class was already in existence, and even before this the
majority of the legionarieswere men whose origin was not recordedon their
tombstones,and who bear colourlessnamestypical of soldiers.I58
Despitethe increasingrole of the hereditarymilitary classthe second-century
army had a socially definablecharacterevenif it is somewhatdifficult to grasp.
The first point for considerationis that althoughuntil about the last quarterof
the secondcentury the auxiliaries were already Roman citizens for the most
part, and therefore were being given diplomas in decreasingnumbers, they
originally camefrom the native populationof peregrinestatus;and as far as can
be judgedfrom the placeswhere their diplomaswere discoveredsomeof them
returned there after discharge.I59 Others remained near their forts, possibly
becausetheir sonswere still servingthere.What the criteria werein the selection
of auxiliary recuits is, of course,impossibleto say, but it is very unlikely that
they were descendedfrom the tribal aristocracy.In the secondcenturythe latter
were probably in the main already Roman citizens, so that their sons had no
reasonto join a unit in order to acquire citizenship through a long term of
157
service.Henceit must be assumedthat auxiliarieswererecruitedfrom the lower
classesin the civitates.
Probablylegionariesrecruitedfrom a colony werealsofrom the lower classes.
The veteranssettledtherewere probablyrich enoughto keep their sonsout of
the army. Sincewe haveto admit that propertyin the coloniestendedto become
concentratedin the handsof the few, this meansthat therewere also poor fami-
lies who could supplypart of the legionaryreserves.The othergroup of soldiers
from the coloniesonly acquiredcitizenshipon joining the legion, and in conse-
quencethey were classifiedas incolae of peregrinestatus;in other words, they
were the native inhabitantsof the territory of a colony and were socially no
higher than an auxiliary from a civitasperegrina.
To sum up: the social compositionof the army may be regardedas decidedly
rustic and plebeian,with the reservationthat many veteransremainedin the
vicinity of the camp and their sons provided part of the reserve. That the
canabaecontinuedin the secondcentury as a special form of community, and
indeed now with regulatedorganization,was probably due to the need for
recruits. The earlier form of organization in these communities was called
veteraniet cives Romani; emphasison the veteranswho were also Romancitizens
can be explainedby the fact that they had a specialrole. The governmentwas
probablyawarethat a group of men uprootedfrom their birth placeand closely
linked to the army could more reliably meet the annual demandfor recruits
than levies in a community where the recruits' background, of necessity
doubtful, could lead to incalculableconsequences, both politically and socially.
The hereditarycharacterof military service establishedin practice, though
without legal regulation,produceda decidedlymilitary classin thefrontier zone,
but onewhich was not in any way isolatedfrom or independentof the provincial
population. The richer membersof this class often becamedecurionsin the
neighbouringtowns, and on the other side the poorer sections of the civil
populationalways had the option of improving their lot by joining the army.
Thus the army provided the link betweenthe two classesof the populationon
the frontier. If we speakof a societywith a military bias or even of a military
societyin the frontier provinces,we have this intermediaryrole of the military
in mind. The results of this development,which beganin the secondcentury,
were clearly evident in the courseof the next.
Given these conditions there was probably no 'veteran problem'such as
occurred in the late republican and early imperial army. The veteranseither
returnedto their birth places,which were probably not too far away, or they
settlednearthe camp; the only problemis that of their role in a civitasperegrina.
As Roman citizens they occupied a privileged position there, but were not
IS8
membersof the native aristocracy,who were citizens by birth and obviously of
long-standingwealth and highly respected.A comparisonof the wealth of the
Eraviscanprincipesand decurionswho could afford wagon-burialswith that of
auxiliary veteransleads to the conclusionthat relations betweenthem and the
aristocratic citizens by birth were strained. Under such circumstancesit was
probably better for the veteransto settle near their former garrisons,I6oeven
thoughliving conditionstherecould not exactly be describedas Romanized.
This leadsto the complexquestionof living conditionsin the first and second
centuries,to which therecan only be a tentativeanswerin view of the dearthof
reliable data. Suchas exist are pointers,ratherthan sourcesupon which definite
conclusionscan be based.The reasonfor this is that sometypes of settlement
assumedto exist have not yet beenexcavated,while earlier excavationswhich
often revealedextensiveresidentialareaswere unsatisfactoryfrom an archaeo-
logical view-point. They failed to recognizethe buildings of wood or clay. Nor
were thosewho carriedout suchwork able to distinguishthe different building
periods.Moreover,theseearlier excavationswere confinedto urbansettlements
and to the living and bathingquartersof the larger villas, and thus providedno
clue to the rural forms of settlementor to those of a definite village character.
Most of the resultsof more recentand archaeologicallybetter-organizedexcava-
tions have either not yet been published,or the work has been restrictedto a
fairly small area,sothat,asfar as forms andtypesof settlementare concerned,we
can only makeassumptions.The towns will be consideredfirst; investigationof
someof themhasbeengoing on for a long time and is approachinga solutionto
someof the most important problemswhich they raise.
The towns and their archaeology

On the basisof what is so far known, and also from the point of view of town-
planning,a distinction must be madebetweenthe coloniaeand themunicipia. At
present informationconcerningthe streetsystemand theorigin and growth of
Flavian and Trajanic coloniae is lacking, but the so-calledHippodamianplan of
the coloniaeat Emonaand Savaria(Fig. 13, p. 75) makesit reasonableto assume
that the otherswereplannedalong the samelines. However,the groundplansof
the municipiawhich haveso far beenexploreddiffer somewhat.The easternhalf
of Aquincumand a slightly smallerpart of Carnuntumhavebeenexcavated,and
thereis a certainamountof information availablefrom ScarbantiaandBassiana.
Aquincum161 had an irregular, trapezoidresidentialareawith a west-eastlength
of some650 m and a north-southwidth of about440 m (Fig. 28). Somewherein
the middle ran the aqueductfrom north to south: it was supportedon piers and
159
-Bat h
-Macellu m N
^ M i t h r a me u
Q so
BathX metres
. Aqueduc t
0 Collegium F
Centonarloru m L
i
"Di
« Sout h v "eDr
si voer jMithraeu m
* Gate ? ri us o
riu
" D i v e r s omr i u
Figure 28 Detailed plan of the municipium of Aquincum

suppliedthe fortressfrom the springsof the so-called'Romanbath', which are
still in usetoday. Excavationswest of the aqueductwerefirst starteda few years
ago, while to the east of it a section some 200 m wide has been completely
exposed.Two virtually parallelstreetslinked by narrowalleys ran on a west-east
alignmentthroughthis part of the town. So far no streetrunning throughthe
whole town from north to southhasbeendiscovered.Hencethereis no evidence
that the Hippodamianprinciple, accordingto which a grid of streetsran in a
straightline acrossthe town from one boundaryto the other, was appliedhere,
althoughthe streetson the whole werelaid out in the form of a rectangulargrid.
The town-centrehas not yet beendiscovered,althoughthere was a commercial
centre,where public affairs were also conducted,roughly in the middle of the
residential area east of the street which ran northwardsfrom the south gate.
Shopsandbusinessesoccupieda colonnadealongthe west sideof this streetand
includedshopsselling Rhenishand Gallic sigillata pottery. On the eastsidewas a
macellum(food market), consistingof a squareperistylewith shopsand a round
building in the middle (PI. 16a). North of it was the so-calledlarge bath-house
(PI. 17). Two public bathshavebeendiscovered,but apartfrom shrinesthereis
no definite evidenceof otherpublic buildings. The collegiumcentonariorumwith its
famous organ was discoverednear the south gate.
The private housesto the south and south-eastof the macellumwere biggish
buildings with an inner courtyard round which a single house or several
togetherformed a roughly squareinsula; in the northernpart of the excavated
areawere narrow houseswith a north-southaspectand mostly separatedfrom
one another by narrow alleys. Small workshops producing metal, leather or
pottery goodswere found in thesehouses;it is possiblethat similar workshops
alsoexistedhereandtherein thosesouthof the macellum.Outsidethe town-walls,
near the south gate, there was an inn (diversorium) with a big yard and stable.
Which of thesebuildings belongto the earliestperiodis not yet clear. The street-
system suggeststhat the settlement developed gradually without any pre-
conceivedplan. Supportfor this assumptionis provided by its earliesttraces,
which date to the Flavian period, approximatelyhalf a century before the town
itself was founded.
The town-walls have beenlocatedon the north, southand westernsidesand
show every sign of having beenbuilt under Hadrian, or at least in the second
century. They encloseda fully built-up area which-as far as can be judged
today-wasnot completeduntil after the walls were built. On the southern
perimeter the outer walls of the houses follow the line of the town-walls,
whereasthe housesthemselvesconform to that of the town sector.The public
utilities date at the latest to Hadrian. The aqueductwas not built for the town
161
but for the legionaryfortress, though the former was allowed to use its water.
Among the recentlyexposedwell-headsat the so called'Romanbath'weresome
votive stones,one of which had been donatedby a decurio municipii et decurio
canabarum. The drainagesystemfollows the twists and turns of the narrow
streets,which suggeststhat it was addedlater. The amphitheatre,a small earth-
bank one, with enclosureand retainingwalls of stone,was completedin its final
form at the latest under Pius, to judge from an altar dated 162 in its shrine to
Nemesis.The excavationswere carriedout more than eighty yearsago and did
not of courseestablishwhetherthis amphitheatrehad beenbuilt on the site of
an earlier one madeof wood.
The cemeteryof the municipiumlay to the west of the town on a road which
linked it to the diagonalAquincum-Brigetioroad. The tombstonesand burials
in this cremationcemeterydate its beginningto the pre-Hadrianicperiod; for
instancethe AgrippinensesTransalpini contributedthe sum of 72 denarii towards
someof the funerals (Fig. 21, p. 125). It is also possiblethat a veteranof legio
X Geminawas buried there before the time of Hadrian (PI. 5C).
To the eastof the town a largepotteryquarterhasbeenexposedwherecrafts-
menwereat work from the mid secondcenturyif not earlier. We shall returnto
this below.
Aquincumhasbeendescribedin somedetail since therehasbeenlittle excava-
tion at other municipia. While it is true that excavationsin the so-calledzoo-
logical gardens(Tiergarten)at Carnuntum162 haverevealeda largesectionof the
town, they haveunfortunatelynot yet revealedits over-all plan. Just beforethe
outbreakof World War II a start was madeon the excavationof a large public
building north of the gardens(the so-calledpalaceruins) and the work is still in
progress(Fig. 29)' In the fifties a large section of the town south-eastof the
Palastruinewas expertly excavated.Thesethreesitesare a considerabledistance
apart.If a roughly east-westrectangleis drawn connectingthe palaceruins and
the part of the town to the south-eastof them (which is essentiallyon the same
orientation)with the remainsin the Tiergarten,this will give an areaof some950
by 600 m (roughly 57 ha), which considerablyexceedsthe size of Aquincum
(some 29-30 ha). This is not at all surprising as the colonia Savariaoriginally
covered40 ha and as early as the beginningof the first centuryCarnuntumwas
alreadyone of the biggesttowns in Pannonia.It is possiblethat the rectangle
just mentioned may have correspondedto the area of the municipium, an
assumptionsupportedby the fact that the so-calledpalaceruins are centrally
situated, while the very simple housesexcavatedin the fifties are round the
edgesof therectangle(PI. 18). Moreover,thesehousesresemblethe free-standing
houseson the northernperimeterof Aquincum, at least in so far as they both
162
The first age 0 fp rosnerity
'.t'
SIRMIVM
\
\
\
\
\
I \
I
I ~Bath
I
I ~ Horreum
I
'PaJ",t",;ne'
Pa.lace? ,,{
o 100 200 300
metres
CARNVNTVM
'PaJ",t",;ne'
Exc.avation~
'Tiergarten
New
Excavations.
Figure 29 Town plans 2

stoodon long narrowplots. The streetsin this part of Carnuntumare not all the
samewidth; two of them runningparallel in a west-eastdirection are linked by
two others which are not quite parallel and do not continue south. Thus this
roughly rectangularstreet-griddoes not conform to Hippodamianprinciples.
Somewherenear the south-westcorner of the rectanglewas the amphitheatre
(PI. 19a), which was probably built in the mid secondcentury if not earlier.
Whether there was a town-wall has not yet been established,but it is very
probable.Whereit could havestoodis still an openquestion,sincecertainindi-
cationssuggestthat not all the built-up areasof the town were in simultaneous
occupation.163
Thereis no doubt that the palaceruins were thoseof a public building, but it
was not possibleto establishjust what its doubtlessvarious purposeswere. The
discoveryof a base164 belongingto a dedicationdonatedby a collegiumin the long
narrow courtyardof the southprecinct suggeststhat there may well have been
an assemblyroom for the collegium within the building, but the whole lay-out
with its staterooms,hypocausts,etc., makesit hardly likely that it was nothing
more than such a centre,even if all the collegia of Carnuntumhad beenlodged
there.
The attention of archaeologistshas been drawn to some of the housesin
Carnuntum,as two neighbouringdwellings therereveala markedsimilarity to a
villa in the Leitha district, at Winden am See,and this type was probablyalso to
be found in Aquincum.I65 It had a colonnadeor arcadeon its narrowfront, and
a corridor runningthroughthe houseanddividing it into two equalparts.There
are numerousexamples,evenfrom pre-Romantimes, of this type of housewith
its T-corridor, but whetherit goesback to local housesof pre-Romandateit is
impossibleto say. Nor would it be right to regard it as the typical Pannonian
house-form,as out of at least thirty alreadyexcavatedin Aquincum only two at
most are of this type, and thereare only one or two examplesamongthe villas.
As for the rest of the housesat Carnuntumthey are more closelyrelatedto a type
found at Aquincum-andone difficult to define. These houseshave hitherto
resistedall attemptsto classify them; their one commonfeature is their irregu-
larly shapedinner courtyard-a squareor rectangularone is a rarity (Fig. 30).
The striking thing aboutthesemodestlittle housesis that they occupyvery little
ground. As the outer rooms were built first and the interior onesaddedlater it
was inevitable that the inner courtyardwas irregular in shape.If the plot was
very narrow, then the courtyardwas either omitted or becamean insignificant
rather than a dominantfeature. Thesehousesshould not be regardedas repre-
sentativeof one or more types.
It was the poorersectionof the communitywhich lived in thesesmall houses,
164
9 .20
, "'10
I
metres
Figure}o Detailed plan of part of the municipium of Carnuntum

The first age ofprosperiry
which also usually servedas workshops.The richer memberslived in houses
which were similar in style to the villas. A small numberof suchhouseshasbeen
excavated,as for examplethe so-calledlargehouseat Aquincum,with its central
courtyard,which, in the final analysis,hadits origin in the atrium of the classical
Romanhouse(pI. I 6b).
Of the other municipia only Neviodunumis under excavation,I66though a
plan is not yet available.The part investigated,in the southof the town, lay on
the banksof Korkorasandwas the businessandharbourarea.On the otherhand
we havea reasonablyclearpicture of the lay-out of the Hadrianicmunicipiumat
Bassiana;althoughit has not yet been excavated,this emergesfrom a highly
successfulaerialphotographwhich is uniquein our region.167 Therewas a single
west-eaststreetrunningright throughthe town, andsomerunningnorth-south
but only through the northernor southernhalves. Somebuildings away from
the town-centrewere not sited in accordancewith this lay-out, henceHippo-
damianprincipleswerenot observedhereeither.It is, of course,not known how
many of the buildings and streetsare late Roman or early Byzantine in date.
Bassianahad a certain importanceup to the end of the sixth century (civitas
Bacensis),but it is very unlikely that in siting later buildings the streetlay-out
was disregarded.
The areaenclosedby the town-walls wassomewhatoval in shapeandits south
and eastsectorswere either not built up at all or the buildings were of clay or
wood. Evenin the town-centreit would seemtherewere someopenspaces.It is
probablethat the town actuallyconsistedof a loosemassof housesandthat only
the centrewhere theofficial buildings were had an urban appearance.Bassiana
was, of course,one of the smallesttowns (c. 19 ha) and, lying as it did between
Sirmium and Singidunum,was an administrativeratherthan a communications
centre. The sameis probably true of the small municipia in the interior of the
province. In the first place thesewere often such small settlementsthat even
establishingtheir sites posesproblemsfor moderninvestigators(Mogentiana),
and secondlyneitherin the Middle Ages nor in more moderntimes were their
sitesregardedas suitablefor the founding of towns. It is alsoprobablethat these
municipia were mostly open settlementswhich did not acquirewalls until the
third or fourth century. Thoseof Scarbantiacontinuedin use and were rebuilt
I'
in the Middle Ages (PI. 2ob) for the town now known as Sopron(Odenburg)
and are still standingtoday; recent excavationshave establishedthat they are
Roman in origin 168 and were only built towards the end of the third or the
beginning of the fourth century. Moreover, not all the buildings were within
their circuit but only the denselybuilt-up area in the centre of the town, the
continentiaaediftcia. Since the beginningsof Scarbantiago back to Tiberius it is
166
BASSJANA
BASSJANA
o, ,
200
I
metres
SCARBANTIA BASSJANA
BASSJANA
A BASSJANA
JAN
S BASSJANA
BASSJANA
S
BA
Kostal
AVRELIANA
Figure]I Town plans 3

not surprising that by the third century it had becomea considerabletown
surroundedby scatteredsettlement,and obviously larger and more important
than Bassiana.The walls in both towns are markedlysimilar in that they enclose
an irregularpolygon, roughly oval-shaped.Thoseat Bassianaprobablyalso date
to late Romantimes.
If the walls in these two towns are comparedwith those of Emona and
Aquincum, which enclose a more or less rectangular area and were either
certainlyerectedat the time the town was founded(as at Emona),or areassumed
to datefrom that time (as at Aquincum), then the assumptionmay be madethat
walls enclosingrectangularareaswere built at the sametime as the towns were
founded, whereasthose enclosingirregular polygonal areasare of a later date.
Thus it is not impossiblethat the walls of Ulpianum,I69 which enclosea large
rectangularareaof some 500 by 500 m, were built under Hadrian.
Sincesettlementswhich receivedchartersevenunderthe Flaviansor Hadrian
did not necessarilyhave an urban appearance,it cannot be assumedthat there
were towns which did not receivethis statusin the first or secondcenturies.The
only exceptionsare the canabaeof the legionaryfortresses,whereurbandevelop-
ment began with the establishmentof the camps; but they were in no way
inferior to the municipia in the secondcentury. As far as their area was con-
cernedthey werebiggerthanthe correspondingmunicipium;on the other hand,
excavationsat Aquincum suggestthat the populationwas not evenly distributed
over theselarge settlementareasand that there was evenless conformity in the
type of house. Large houseswhich indicatedno small degreeof luxury stood
cheek by jowl with small primitive shacks and workshops, and there were
probably more buildings made of wattle and daub than in the neighbouring
municipia. This does not, however, mean that the canabae had none of the
communalservicesindispensableto a Roman town. Some of the streetswere
paved and seweredand some parts of the canabae were definitely urban in
character.The urban part of the canabaeat Viminacium,I70 investigatedin the
last century, had not only pavedand seweredstreetsbut a colonnadealong one
of them,while thoseat right anglesto oneanothersuggestsomedegreeof town-
planning. This is by no meansimprobable,as the canabae,at least sincethe time
of Hadrian,hadpossessed not only decurionsbut also all the magistratesrespon-
sible in an urbancommunityfor local affairs. Apart from the big market-placeat
Carnuntumno public buildings were discoveredin the canabae,but this may well
have been fortuitous. The so-called military amphitheatreson the perimeter
of these settlementswere probably purely military structures. A building-
inscription of the time of Antoninus Pius which mentionslegio II Adiutrix is-
in view of the placewhereit was found-no doubt rightly taken to refer to the
168
military amphitheatreat Aquincum. Civilians from the canabaewere also prob-
ably admittedto the gamesheld in the amphitheatres.Thoseat Carnuntumand
Aquincumwerebiggerthanthe onesbelongingto the municipia, andthe sizeof
the caveas indicate that they were designed to accommodatea substantial
number of spectators.171
Nowhereis thereevidence,nor is it probable,that the canabaewere walled. It

would haveviolatedall the principlesof tacticsif the fortresshadbeenprotected
by a secondwall which encloseda very large numberof civilians.
Villas and rural settlements

Whereas some information is available concerning the history of urban
settlementsthis is far from being so with rural ones, so that even a tentative
description of them is out of the question. All that is possible at presentis
to indicate some of the house types, so that by putting them into categories
and establishing their geographicaldistribution certain inferences may be
drawn.
Villas which can certainly be datedto the time of MarcusAurelius are few in
numberand found only in someparts of the province. 'Villa' is here taken to
mean housesor groups of buildings which either formed part of an evident
agriculturalcomplexor which, in default of investigation,may be assumedto be
such.Seeminglydetachedbuildings may well havebeensmall housesin a village
settlement,parts of staging-postsor severalother possibilities. The important
thing is that housesbuilt in the techniqueintroducedby the Romans(stonewall,
mortar and roofing tiles), but which cannot be regardedas villas, have so far
beenfound only in the third and fourth centuries.Thus Romanbuilding tech-
nique seemsto have been restrictedin rural areasto villas and to what were
obviously military and official buildings; it was only later that it spreadto
settlementsof non-villa type.
In Pannoniavillas dating to the early imperial period havebeendiscoveredin
thoseareaswhereforeigners,mainly north Italians, settledin the courseof the
first centuryor elsein areaswhere,althoughthere were no externalsettlers,the
native aristocracycameto terms at an early date with Romancontrol-a com-
promisewhich was then sealedby the grantingof citizenship.This meansthat so
far early-imperial villas have beenfound only on the westernbordersof Pan-
nonia, in the Leitha district (Boians) and north of Lake Pelso.At the sametime
it should be borne in mind that there has beenso little excavationbetweenthe
Drave and the Savethat the fact that no villas have beenfound therecannotbe
used as an argument. The situation is entirely different on the northern and
169
Donnerskirchen
Parndorf ~ Winden
"/ am See
Regelsbrunn .
Eisenstadt ,
o 100
, :;wo
I
metres
Figure J2 Villa plans I
170
easternbordersof the province where neither earlier nor more recent excava-
tions of non-urbansettlementshave revealedvillas of the early empire.
A clearly distinctive type is representedby villas in which the domestic
quartershad a central courtyard(Fig. 32), probably unroofed.I72In somecases
this type of courtyard, which owed its design to the classicalatrium, had an
impluvium. It is not necessaryto regard therooms groupedround the small
courtyardas alae,tablinum,vestibulum,fauces,etc. That suchpedantryis out of
place becomesclear from the fact that the entrancewas not normally on the
narrow side of the building oppositethe main room with the apse. In some
instancesthis apsidal room was missing or had been added in a haphazard
fashion,as at Eisenstadt.The bathroomswere eitherin a separatebath-houseor
in the houseitself. The deciding factor was the size of the whole complex.
In view of the small number of ground plans available it is inadvisableto
attemptany further differentiation of types, as it is not possibleto distinguish
betweena type and an individual case(Figs 32, 33). Among 'types'we may place
big palaceswith two central rooms,173long rectangularhouseswith a suite of
roomsand onelargeinterior one,I74·the housewith aT-corridor,etcP5A more
importantclassificationis that basedon the size of the domesticquartersand of
the whole complex.In a few casesthe farm buildings andthe walls enclosingthe
rectangularyard havebeenexposed.Thesewalls were not a defensivemeasure
as they are mostly very narrow and there is no evidencefor fortified gatesor
ditches. Their purposewas probably to protect the yard where corn, produce
and implementswere kept and in particular to keep stock from escaping.The
housewas cut off from the yard by an additional inner wall as a protectionfor
the orchard,the kitchen gardenor evenjust the flower garden,or perhapsit was
to provide the residentswith quiet and privacy.176
The biggestcomplexeshad a yard measuringsome200 by 200 m, or possibly
more (Balica, Parndorfand Smarje) with a houseno less than 30 by 40 m; in
everycasethe bath-housewas separate(Fig. 32). The farm buildings,it hasbeen
established,were along the yard-walls and are only rarely to be found in the
middle of the yard; what thesebuildings were usedfor it is difficult to say with
certainty, but there are instancesof granaries(horrea) having beenfound with
more or less typicallay-out, and as they are of a striking size they point to the
existenceof very large estates.The villa at Parndorfhasthe biggestyard. house,
bath-houseand granary. It is in the territory of the civitas Boiorum and was
built towardsthe endof the first century.It is probablycorrectto assumethat it
was the seatof a very rich memberof the Boian aristocracy.
The smallervillas havea rectangularyard, some90 by 100 m andmostly small
houses(roughly 10 by I) to 20 by 2) m) which also contain the bath-rooms
171
BaJaca.
BaJaca.
BaJaca.BaJaca.
o, 100
, ,
zoo
metres
Figure JJ Villa plans 2

(Fig. 33). Such measurements madeit possibleonly in rare instancesto build a
housebasedon the atrium type (the peristylevilla): a centralcourtyardrequired
a larger complex.
The ethnic and social backgroundof the ownersof theseearly-imperialvillas
is fairly clear. Thosein the upperDrave valley, the Pelsoregion and somenear
Scarbantiabelongedto foreign settlerswho hadmanagedto acquireland andhad
introducedthe Roman method of farm management.In the Drave valley the
latter probably started with the veteran settlementsmade at Poetovio under
Trajan,but nearScarbantiaveteransandnorth Italian tradershadalreadyformed
a corporation in the time of Tiberius and in the Flavian-Trajanic era were
establishingthemselveson the shoresof Lake Pelso.In the Leitha region, where
both large and small villas have beenfound, it was obviously Boian landowners
who adoptedthe Romanway of life andits farming methods.SomeBoianswere
very rich, others of more moderatemeans.It often happenedthat the smaller
estateswere mergedwith the bigger ones;evidenceof this was found in the fact
that despitethe small numberof villas so far excavatedsomewere not inhabited
after the end of the secondcentury,I77 whereasthe huge villa at Parndorfand
other large establishmentswere occupiedright into the late fourth century and
show signs of having beenconsiderablyaltered.
The villa with the T -corridor at Winden am See was in the territory of the
civitas Boiorum. As far as presentknowledgegoesthereis no justification at all
for seeinga connectionbetweenthis type of houseand one presentlocally in
pre-Romantimes. Unfortunately it is not known how or where those who
workedon the Boian estateslived. In otherparts of the province,however,some
late La Tene and early Roman houseshave been excavatedwhich representa
completely different type from that with a T-corridor (Fig. 34). To attempt a
generalizationfrom datanow availablewould be premature;thereis, however,a
type which standsout as it occursboth in late La Teneand Romancontexts.In
the oppidum on Mount Gellert at BudapestE. Bonis explored some roughly
rectangulartimber houses,with rounded corners, occupying shallow excava-
tions in the soil; the hearth or the oven-whichwas made of clay and sealed
when in use-wasinside the house,near one of the long walls. Thesehouses
were at least4 m long and 2-3 m wide. The roof was supportedby two beams,
which stoodin the centreof eachof the shortersides.I78 Similar housesfrom the
secondcentury have been discoveredat Gorsium,I79 and the type is known
beyondthe frontier in the Roman Iron Age of Slovakia and Moravia.I80 The
civil settlementattachedto the auxiliary fort at Matrica (Szazhalombatta)also
consistedin the secondcenturyof similar dwellings, althoughthe groundplans
have not beenestablishedin detai1.I8I
173
post-hole post-hole
post-hole
Late Celtic House post-hole

Mount Gellert ( Budapest) , _...i'==:::iT
metre's 0..
2ndcentury Houses
post-hole : metres o7 5'
',,
70 15
post-hole
post-hole
2ndcentury HousesHouses
2nd century
Tac metres
o7 5'
',,
70 15
Figure J4 Plansof native houses
174
To dateno houseswhich werenot below groundlevel have beendiscovered.
The probablereasonfor this is that they arevery difficult to trace.Walls of terre
piseeor timber soondisappearcompletelywheresurfaceconditionsare the same
today as they were then. It is only at Halimba that groundplans can be recog-
nizedof huts, madeof woodandclay.I82If this interpretationis correct,we have
a further type of native house(Fig. 34). On the basis of the ground plans at
Halimba it was small and either round or square,with a single entrance;the
former hada dhtmeterof 4-8 m, while the latter was 5-7 m across.Their dating
approximatelyto the period from Hadrian to Marcus is attestedboth by loose
finds from the siteandalsoby gravesin the vicinity containingcrematedremains.
Roundandsquarehuts arealso attestedfor the sameperiodoutsidethe frontier;
it is only necessaryto draw attention to the reliefs on the column of Marcus
Aurelius.
The discoveryof thesehouseshas,however,not broughtnearera solutionto
the problem how and where the native aristocracyin northern and eastern
Pannonialived. In the first placeit is hardly likely that aprincepsor a decurio who
could afford to have an extravagantwagon-burialand indulged in gymnastics
(palaestra-equipment)lived in such miserable pit-dwellings. An auxiliary
veteranmight perhapsbe contentto live in a bigger housebuilt in this manner,
since,after all, the housesin the auxiliary civil settlementsof the secondcentury
were hardly better.I83 The discoveryat Aszar of a hoardof somebronzevessels
and simple piecesof silver jewelry (PI. 21), belongingto an Azalian called Atta
Nivionisfilius, accordingto a diploma(PI. 22) found with them, doesnot indicate
any greatwealth on the part of the owner.I84The tribal aristocratswere much
richerandmoreexactingin their demands.Perhapssomeindicationof their mode
of life will be providedby the finds at Gorsium(Tac), where,accordingto local
inscriptions, some native decurionsof Aquincum lived.I8 s It was in Gorsium
that an elaboratetombstoneof importedcrystallinelimestoneand of first-class
designwas set up to a decurion. But the buildings which can be dated to the
secondcenturyso far excavatedthere-thoseof stoneare mainly late Roman-
belong to the probably once very impressiveshrine where the Ara Augusfi
Provinciae Pannoniae Injerioris stood, and where the Concilium Provinciae met
regularly. This shrinewas called TemplumProvinciae (Fig. 41, p. 257; PI. 23a).I86
Thereis probablylittle hopeof discoveringany greatnumberof estatecentres
belongingto the eastPannoniantribal aristocrats,sincethe latter were not very
numerous.Nor is it known whetherand to what extentthey adaptedtheir farm-
ing methodsto Romantechniques.If there were no such change,then, while
therewould be housesand estatecentres,we cannotexpectto find villas. It was
preciselythesearistocratswho indulgedin the traditionalandextremelyextrava-
175
gant wagon-burials,whereasthe Boian aristocratswho lived in villas had no
such custom. Native dress and names persistedmuch longer in north-east
Pannoniathanamongthe Boians,wheretombstoneswith Celtic namesandwith
the deaddepictedin native dresswere becominglessfrequentin the first half of
the secondcentury:in north-eastPannoniathey persistedin largenumbersright
up to the Marcomannicwars. The merefact that nativesset up tombstoneswith
Latin inscriptions by no means implies that they lived in stone houseswith
hypocaustsand bathrooms; in north-eastPannoniathis Roman standardof
living can be assumedonly for the richestnative aristocratswho were buriedin
wagon-gravesalong with equipment from the palaestraand other Roman
luxury goods.
Industrial development:the pottery industry

We cannotassumethat a radical changetook place in the living standardsof
the broadmassesfrom the merefact that everywherein the secondcenturylocal
industry wasin a position to meetthe demandfor goodswhich, both in design
and craftsmanship,were becomingincreasinglyRomanized.In the first place
the late La Tenepotteryandbronzeswerenot technicallyinferior to the Roman,
and, second,the spreadof Romandesignswas largely a matterof their being the
latest fashion. 'Roman' goods have been found in settlementsin north-east
Pannoniaalongsidethose basedon local tradition, regardlessof whether the
settlementwas a purely civilian one in the hinterland of the limes or was an
auxiliary vicus. By the end of the first century Roman designshad gained the
upperhandin westandsouthPannonia,but in the north-eastof the provincelate
Iron Age designswerestill beingwidely usedin the secondcentury,particularly
in pottery in the Eraviscanarea (the so-called 'grey ware of Patka').I87The
pottery industry in Pannoniacan be tracedback clearly to that of the late Iron
Age,I88 the influence of which in design and technique persistedinto the
Romanperiod. From that time until late antiquity pottersusedthe sametype of
kiln which had alreadybeenin use during the Iron Age. The so-calledRoman
kiln, the square, walled type,archedwith tiles, was muchrarerandprimarily used
for firing tiles.I89
The most characteristicproductsof local potters were vesselswith stamped
decoration (PI. 24a); these were made and used throughout the whole of
Pannonia.I90 From the point of view of techniquethey derive fromthe stamped
pots of the late Iron Age. The earliestwest Pannonianexampleswere, however,
imitations of north Italian disheswhich hadthe potter'sstampin the centre.The
shapeof thesevesselswas imitated by the local pottersbut given a simple leaf
176
stamp.Around the end of the first centurya Celtic potter,perhapsfrom Aquin-
cum, stampedhis name RIISATVSFIICI (Resatusjeci) on similarly shapeddishes
(PI. 24a); in the first half of the secondcentury pottersstartedimitating Gallic
samian.Insteadof disheswith leaf stamps,vesselsof Dragendorfform 37 were
producedwith the relief decorationreplacedby stamped decoration(PI. 24a).
Theselocal imitations of samianwere widely distributedthroughoutthe whole
of the secondcenturyand probablycontinuedinto the Severanperiod, particu-
larly in thoseareasinto which genuinesigillata had not yet penetrated.
As far as other kinds of pottery are concernedit is not always possibleto
distinguishbetweenimportedgoodsandthoselocally produced,sinceamongthe
so-calledPo valley andlate north Italian sigillata therewas a simpler shapewith
a poor glazewhich may well havebeenmadein Pannonia.I91 It is very unlikely
that massesof simple glassandpotteryvesselswereimported into the province
when they could have beenproducedon the spot and so have savedhigh trans-
port costs.Mass imports can only be attestedfor brief periods,as for example
during the Flavian-Trajanicperiod,whenthey weredestinedfor the Danubegar-
risons.As soonas local industrycould meetdemands,it was alsoable to oustim-
portedarticles,taking over their designsand methodof production.From then
onwardsimports were restrictedto certainspeciallines such as samianpottery,
which were,however,distributedonly along the main routes,in particular the
waterways. Thus in the more remote areas greater importance attached to
imitations, as for exampleto samiancopies with stampeddecoration.I92
Evenlocally producedsamiandid not reachareasoff the beatentrack. One of
the oddestfeaturesof Danubiantrade and Pannonianindustry is that whereas
potterswereproducingsigillata the provincewas hardly awareof the fact. So far
two potteriesproducingsigillata have been discovered-oneat Aquincum and
the other the so-calledSiscianpottery. On the Danubebank eastof the munici-
pium at Aquincum, excavationshave revealed an extensive pottery quarter
where domestic pots, tiles, terracottasand samian were produced.I 93 Large
quantitiesof punches,mouldsfor bowls, and othertools usedin makingsigillata
cameto light and madeit possibleto distinguishtwo craftsmen,the so-called
'first master'and the potter Pacatus(PI. 2p). The typical decorationsthey used
were vine leaves,vine tendrils with hand-incisedstems,hares,dogs, wild boar,
figures of Panand a peculiarfigure of Vulcan. On the basis of the well-known
moulds, which have beenfound in such large numbers,it should be easy to
identify the products among the finds in the settlements,but none has been
found in Pannonia.Thus it is all the more remarkablethat a bowl mould and
somesherdswerefound at Mursaand a fragmentof stuccowith Pacatus'stamp
on it at Tricornium (Ritopek), east of Singidunum.I94 This suggeststhat the
177
Pacatusworkshopeither exportedalong the Danubeto Moesiaor was attempt-
ing to do so. The mould found at Mursais evidenceof the regulareffort to move
the pottery nearerto the market. The existenceof the so-called'Siscia' pottery
is known only throughthe finding ofits productsin southPannoniaandMoesia.
L. Nagy~who identified this pottery (PI. 24b), suggestedits location at Siscia;
but it could equally well have beenin someother town in the Savevalley.I95
Both potterieswerein operationroughly in the dec-adesfollowing the middle
of the second century,that is to say at the very time when Lezoux, Rheinzabern
and Westerndorfwares,to mention only the most important, had practically a
monopolyin Pannonia.Local potteriescould not drive them from the market,
but were probablyable to competeon the lower Danube,wheretheir transport
costs were lower than those of the potteries of the west. It is probably also
significant that at this time local demandcould not be met locally,although
Pacatuswas turning out a fairly successfulproductwith a good ghze. Thus the
conclusioncannotbe avoidedthat the price differencesbetweenGallo-German
and Pannoniansigillata, which must have operatedin the Pannonianmarket,
werenot sufficiently greatfor the demandon the part of the masses-whocould
not afford westernsigHlala-to be met from local production. Local products
could probablyonly oustimportsif their quality was equallyhigh or their prices
muchlower. That his potterycloseddownfairly quickly is evidencethat Pacatus
could not producecheaplyenough.It is also probablethat he was unableto win
the Moesianmarket; underMarcusAurelius the pottery eastof the municipium
of Aquincum was producingonly domesticpots and terracottas.
Locally producedwares consistedof vesselsderived from those of the late
Iron Age but influencedby imports, or else of direct imitations of the imports
themselveswhich gradually madethe latter superfluous.Specifically Pannonian
shapesof vesselare derivedeitherfrom the weakerungtraditionsof the Iron Age
or from the forms of importedpottery. Especiallyinfluential amongthose types
which did not go back to the late Iron Age were the onesthat had reached the
province at a critical period either by way of import or in the train of foreign
craftsmen.This critical period always set in when demandsuddenlyincreased
several timesover as the result of mass immigration either of soldiers or of
civilians. In westernPannoniathis occurredat the time of the colonial deduc-
tions and the immigration of north Italians, whereasin the easternpart of the
province it coincided with the building of forts ·on the Danube in the late
Flavian-Trajanic period.
178
Sculpture
What has beensaid aboveapplies equally to artistic activity, if that is the right
word. Just as pottery has beenusedto representthe whole rangeof industrial
production, so stone sculptures,primarily grave sculptures,must serve as an
exampleof 'art'. The only clifferencebetweenthe two is that the native popula-
tion has no tradition in the field of sculpture,and particularly not in that of
funerary sculptures.Attempts to link someof the characteristicsof early stone
sculptureswith pre-Romanart arenot convincing.The only explanationfor the
clumsy and inexperiencedworkmanshipis that the native stonemasonwas not
accustomedeitherto working in wood or to any kind of sculpting. The clearest
exampleof stonemasons' work revealingtraining in wood, or lack of it, is to be
seen in the so-called astral symbols on tombstones(Fig. 10). The whole of
Pannonianstone sculpture goes back to designs introduced in the critical
periods. Since the custom of setting up tombstonesfirst came in with the
Romans,there is an absenceof any native influence, except perhapsfor the
astral symbols. The earliest stonemasonscamefrom northernItaly and intro-
ducedtheir own views concerningcompositionandornamentation(PIs loa, I la,
b). When in some areas the natives began to set up tombstonesthey either
orderedthemfrom masonsin thetownsor, if theseweretoo far away,from native
stonemasons who in the early dayswerecompletelyunskilled.Theseimitatedthe
work of the yards alreadyoperatingin Pannonia,andproducedinterestingand
extremelycharacteristicgroups of monumentswhich both in compositionand
ornamentationreveala direct link with the work of neighbouringstonemasons
(PI. 6a). Suchgroupsin the Leitha district I96 dateto the Flavian-Trajanicperiod
and in north-easternPannoniato the late Flavian-Hadrianicperiod.I97
Until aboutthe time of Hadrianthe 'artistic' output of the stonemasons' yards
was thus of two kinds: the yards which were developingdemandin the big
centreswere working according to the traditions of northern Italy and with
craftsmenwho either camefrom thereor were apprenticedto northernItalians.
At the sametime native stonemasons were working for the nativepopulationin
areasto which the productsof the 'Roman'yardscould not be transported.The
differencewas thereforenot who had orderedthe tombstonebut whereit was to
be erected.From the time of Hadriantherewas a levelling out; the bettercrafts-
men began to open yards wherever there was a demand,and the primitive
productsof local stonemasons disappeared.The 'specifically Pannoniancharac-
teristics'in stonesculptureswere thus designsintroducedby stonemasons from
northernItaly and taken over by their successors.Thus the earliestinfluences
remainedfor ever decisive.
179
Upper Moesian stone sculpturesdevelopedalong the samelines, with the
differencethat primitive imitations by local masonsarefound only very sporadic-
ally becauseof the lack of demand.The earliestworkshopswere in the Flavian
colony of Scupi, where simplified imitations of Macedoniansculpture were
produced.This type of production,to the exclusionof the others,persistedright
into the third century. The workshopswhich were openedsomewhatlater at
Singidunumand Viminacium (PIs 6b, c) were influenced by those in northern
Italy, or to be more precise, by northern Italian designs producedin south
Pannonia,whereasthe sculpturesfrom Ratiaria were partly influenced by the
latter andpartly by the workshopswhich had alreadystartedup at Oescusin the
first century. Thosefrom Ulpianum and Timacum minus are an emanationof
the traditions which had arrived there along the Danube(PI. 26).198
Both in Pannoniaand Upper Moesia the most important characteristicof
tombstonesculpturesfrom both town and country was that once theseinitial
influences were acceptedthey proved decisive, whereasthe specifically local
contributions consisted overwhelmingly of quite unoriginal characteristics.
Thus tombstonesin south and west Pannoniaare tympanumstelae(PI. I I) on
which the spacefor representationeither played a subordinaterole or was left
out altogether; portraits are found only rarely. It is the design itself-the
tympanumon columnsor pilasters-whichcreatesthe dominantimpression.In
northernPannonia,on the other hand,wheretympanumstelaehad also become
customaryto the exclusion of other styles, the stele is coveredwith sculpted
portrait-busts(PIs IIC, I2a). The Upper Moesianstonemasons' work was even
more limited in scope.The tympanum,half-columnsor pilasters,the spacefor a
wreath or portrait, the framing tendril-ornamentround the border, the vase-
and-tendrilsmotif and various iconographicaldesignswere all taken over from
south Pannonia(PIs 5a, 26a). They were, of course,not all madeuse of every-
where,but only a limited numberof themin eachtown, which then cameto have
a few types of its own. For instance,Ratiaria had stelaewith an archedtop, a
wreathwith a rosetteandthe vase-and-tendrils motif, Timacumminus hadthese
with portrait-bustsin addition, while Ulpianum had only busts, the tendril-
frame and the vase-and-tendrilsmotif (PI. 26). The limited natureof the motifs
was so pronouncedthat at Ratiaria and Timacum minus the vase-and-tendrils
motif is cut off from the tendrils on the edgesand given a spaceto itself; at
Ulpianum, on the other hand,the tendril is always shown growing out of a pot
(cf. PI. 26a).This vase-and-tendrilmotif is not found at Scupi, Viminacium or
Singidunum.Originality on the part of the stonemasons thus consistedsolely in
the way they combinedcertain decorativeelements.
It is only rarely that the productsof stonemasons'workshopscan be regarded
180
as works of art. Most of the tombstoneswere mass-produced, as in the caseof
the stelaewith wreathswhich were made at Aquincum199 in the first half and
middle of the secondcentury for the collegia, and were then copied by native
craftsmen(PIs 5C, 6a). It is only at Viminacium and Savariathat there is any
evidenceof artistic quality (PI. 6b, c; IIb). If the citizens of a town wantedto
commissiona work of art of high quality, they sentfor a foreign sculptor.About
the middle of the secondcenturystatuesto the Capitoline Triad were set up at
Savariaand Scarbantia,and were the work of sculptors either from northern
Italy or from Virunum (PI. 27a ). The stone,too, was imported.20o
Templesand the developmentoj religion

The building of Capitoline temples in two west Pannoniantowns must be
regardednot only as a symptombut also as a symbol of economicprosperity.It
was in addition the time when amphitheatreswere either being built or rebuilt
in stone,and when the theatreat Scupi was probably constructed.It could be
said that by the mid secondcentury, in the eyes of the urban aristocracy,the
provincehadbeengiven its finishing touchesand had becomeRoman.This rare
momentin which there was a senseof having reachedmaturity and fulfilment
was, of course,capturedin the RomanOration of Aelius Aristides.
The capitolia at Savariaand Scarbantiaare also important for the history of
religion. In contrastto the west Europeanprovinces where native places of
worship were turned into big Gallo-Roman or Romano-British shrines, in
Pannonia,andprobablyin UpperMoesiatoo, it was to official Romangodsthat
large shrineswere erected.Consequentlywe have hardly any gods with native
namesor at least with native titles, or even possessingnative characteristics.
201
The evidenceof first- and second-centurymonumentsshows that religion in

Pannonia and Upper Moesia was unmistakably Roman. There were slight
exceptionsto this, such as the Egyptian gods worshippedby the urban aristo-
cracy in west Pannonia,especiallyat Savaria(PI. 27b), wherethe cult had been
introducedvia Aquileia.202 Moreoverat an early datethe legions who had been
fighting in the eastunderTrajanandHadrianbroughtbackwith themthe cult of
JupiterDolichenus;the earliestrecord of him in the west is the votive inscrip-
tion set up at Carnuntumby the iuventuscolens Iovem Dolichenum.203 The cult of
Mithras was also introducedat Carnuntumat a very early date;204 the earliest
Mithraeumin Pannonia,at Poetovio,datesto the reign of Antoninus Pius, the
cult being founded by customs officials.20s Thus both the military and the
administrationhad an impact on the developmentof religious belief.
Neither Romanforms nor any local onesthrow light on the religious beliefs
181
of the natives. Up to Severantimes, apart from the goddessAecorna or
Aequornaat Emona,there is no evidencefor any native gods or native cults.
This alsofits in with the characteristicsof Romanizationin PannoniaandUpper
Moesia, evenif it is not possibleto formulate the reasonsprecisely.In the first
placethe conservatismof all religions must be underlined;it was responsiblefor
the retentionof traditionalcustomsin the forms of worship. The nativesdid not
erectaltars totheir own gods-if'god'is theright word here-whileeventhose
they erectedto Romangodsareextremelyrare; theyareprimarily official dedica-
tions, suchas that setup to Jupiterby the duoviri of Aquincum. Thus the native
populationshowedlittle interestin the religion of their conquerors,which is the
reason why the native gods were not identified with those of the classical
Pantheon.Identifications were the work of foreigners, who in the first and
secondcenturieslikewise showedscant interest in the religious beliefs of the
natives.Savus,Danuviusand similar godswho wereworshippedby tradersand
officials 206 were nothing more than the genii loci whom outsiders honoured
whereverthey happenedto be. Thus the reasonwhy it was preciselythe Severan
periodin which native religious characteristicsemergedandweredocumentedis
not hard to understand.Even at that time politics and religion were closely
connected,and as long as the province remainedpolitically of no consequence
its religion too showedlittle individuality. As soon as the situation changed,
however,foreignersandthe local inhabitantswereequallyinterestedin harness-
ing the nativegodsto the serviceof groupinterests.Foreignersgavenativegods
identificationswith classicalones,while the nativessuddenlyshowedinterestin
thoseof the foreigners.
182
Cbapter 6
Crises on the Danube:
the rise of the Illyrican soldiery
'Dum Parthicumbellum geritur, natum est Marcomannicum,quod diu eorum,

qui aderant,arte suspensumest, ut finito iam orientali bello Marcomannicum
agi posset.'('While the Parthianwar was still in progressthe Marcomannicwar
brokeout but was skilfully postponedby the menon the spotso that the eastern
conflict could be finished first. ') This statementis found in the Life of Marcus
in the Historia Augusta,I and there is no reasonto doubt it. The fact that when
the Parthianwar beganin 161 a Marcomannicwar could not yet be foreseen,is
proved by the fact that Marcus withdrew legio II Adiutrix from the Danubian
army and employedit in the Orient.2 It is true that this legion was replacedat
Aquincum by lIIl Flavia,3 but, as we have already seen,this legion was con-
sidereda permanentreservefor PannoniaInferior. If therehadbeenanyimmedi-
atedangerof war on the Danubearound161, Marcusis unlikely to haveordered
one of the most important Danubianlegions to the Orient. Moreover, vexilla-
tions of practicallyall the Danubianlegionstook part in the Parthianwar; 4 this
is sufficient proof that the Danubianregion was quiet. Dangerof war on the
Danubebecameevident only later, when Marcus even decidedto recruit new
legions. It was presumablyto the credit of the governorsof Pannonia,and
perhapsalso of Noricum and Dacia, that the imminent war was postponedby
diplomatic meansuntil the Parthianwar could be terminatedand the troops
which had beensent thither from the Danubecould return home.
The fact that it was possibleto delay the outbreakof the war for someyears
by negotiationsis of importancein determining the causesof the war. The
ars eorumqui aderantpresumablymeantthat the governorsknew how to prolong
negotiationswith the barbarianclient-princes.This, however,was possibleonly
183
Crises on the Danube: the rise oj the II(yrican soldiery
if the client-princeshad made demandswhose fulfilment they could not con-
sider to be wholly impossible.This implies that the negotiationsstartedwithin
the framework of the systemof c!ientela; the governorswere able to raise in
their oppositenumbersthe illusion that their demandswould be consideredand
fulfilled at the highestlevel. Thesedemands,however,were certainlynovel and
unprecedented; otherwisethe emperorwould not havefelt it necessaryto make
seriouspreparationsfor war while the negotiationsthemselveswere still going
on. The recruitmentof two legions-II and III Italicae-wasin itself a measure
that did not fit into the patternof 'ordinary' frontier wars, wagedfrom time to
time in spite of, or becauseof, the system of client-kingdoms. The unusual
demandsof the barbarianscalled forth unusualmilitary measures:this is a fact
which requires particular emphasis.The recruitment of new legions can best
be explainedby an intention of the emperorto establishnew provinces,s and
it is in fact reportedthat Marcus intendedto annex two new provinces:Mar-
comanniaand Sarmatia.6 If our assumptionthat the two new legions were
intendedfor the two projectedprovincesis correct,thenwe must also conclude
that Marcus had already planned the annexationbefore the outbreak of the
war. Although the Historia Augustaat one point attributes the failure of this
programmeto the revolt of A vidius Cassiusand at another to the death of
Marcus-a chronological contradiction-it seemsthat the plans of conquest
had already beenabandonedwhen negotiationswith the barbarianswere dis-
continuedat the outbreakof war. The reasonsfor this conclusionare that soon
after its recruitmentlegio II Italica was given a stone-builtfortress in Noricum
(Locice),7andthat both newlegionshadalreadybeenassignedto their final base-
fortresses, Castra Regina (Regensburg)and Lauriacum, in the last years of
Marcus, as if they were no longer intendedfor newly conqueredterritories.
Presumably,therefore, Marcus had already given up his plans of conquest
c. 166 or 167, and the conclusionmay be drawn that it was the outbreakof the
war that promptedhim to do so. The plans for Marcomanniaet Sarmatia, then,
somehowarose in connection with the negotiationsbetweenthe governors
and the Danubianbarbarians.But how? It is obvious that the nature of the
barbarians'demandsmust be deducedfrom their subsequentbehaviour and
intentions, about which CassiusDio gives some information. It is frequently
mentionedduring the courseof the war that groupsof barbarians'surrendered
to the emperor',or offeredarmedhelp in exchangefor land in a province,or that
they specifically tried to settlein the empire-andthis Marcus actually allowed
somegroupsto do. The barbarianshadprobablyalreadycomeforward with the
samedemandsbefore the outbreakof war, as well as in the negotiations,and
this was apparentlyan entirely novel factor in Roman-barbarian relationswhich
184
Crises on the Danube: the rise oj the Ilfyrican soldiery
might well resultin a novel reactionon the Romanside.At first sight it doesnot
seemvery likely that the barbarians,who during the precedingcenturieshad
takenup arms againstRome severaltimes, and againstwhom a powerful army
had to be maintainedon the bank of the Danube, should suddenly wish to
exchangetheir freedomfor Romanbondage.Weareaccustomedto think of the
role of an army as offensive, particularly in the caseof an empirewhosesize is
somehowlogically connectedwith its continualgrowth; an empiresomehowhas
to be imperialistically orientated.How, then, could the enemiesof the empire
demandtheir own suppression? But in the secondcenturythe Romanarmy was
anythingbut offensive. Contemporarynoticesfrom Florus8 to Tertullian prove
the contrary: the profitability of Rome's conquestswas questioned,the limes
was regardedas the wall of the empire,whosepurposewas seenas not only to
protect the empirefrom enemyattacksbut also to excludethe barbariansfrom
it: quot barbari exclusi! Tertullian exclaims.9 Since the time of Augustus the
elaboratesystemof treatiesconcludedwith barbariannationson the right bank
of the Rhine and the left bank of the Danubeformed an organic part of this
systemof securityanddefence.The 'exclusionof the barbarians'from the empire
apparentlybecameincreasinglya problemof foreign policy, and this can easily
be understoodif we considerthe difficult situation of the barbarianfrontier
people. According to the terms of the treaties made with barbarianprinces,
the contractingparties had to assisteachotherwith arms: but this, of course,
never implied any inclination by Rome to make armed intervention in the
hostilities of rival barbariantribes or princes. The barbarians,on the other
hand, had to maintain peaceand quiet along the border, and in return those
barbarianprincesand their followers who were faithful to Rome were allowed
to participatein Roman wealth and the splendoursof the empire. Raro armis
saepiuspecuniawas the sloganof this policy basedon treaties;10 it led the allied
peoples sooner or later to the conclusion that it would be better to be the
subjectof Romethan her ally.!1
It is clear why this problem had becometopical in the sixties of the second
centuryin particular. Among modernhistoriansit has becomea commonplace
to ascribethe Marcomannicwars in the last analysisto the Gothic migration: the
greatmigrationsand displacementsof peoplesin north and eastEurope,begun
in the first half of the secondcentury,had an impacton the peopleon the border
of the Roman empire, and causeda north-to-southor possibly east-to-west
pressureon the limes.12 This may be roughly correct.But the negotiationswhich
had precededthe war suggestthat thoseDanubianbarbarianswho were allied
with Romehadinitially tried to attachthemselvesto Romeor to win Romeover
for a fight against the barbarianson the move.13 The forces of Rome were,
185
Crises on the Danube: the rise of the Illyrican soldiery
however,tied in the Orient, and for this reasonthe governorsof the Danubian
provinceshad good causeto delay an armedconflict. Their successprovesthat
the barbarianfrontier peopleshad not beenresolvedon war with Rome from
the first. And during the negotiations,the possibility of m'<lking provincesof
barbarianterritory beyondPannoniawas presumablydiscussedtoo; in the exist-
ing situation it could provide a solution welcometo the threatenedDanubian
barbarians.
This reconstructionis confirmedby the fact that the first enemiesto crossthe
limes and attack the empire, as far as we can tell, were not the Marcomanni,
Quadi and Sarmatians,but the Lombards,Obii and perhapsalso the Vandals,
all of whom lived north of them. A hostile movementof the frontier people
themselvesis first recordedonly from 167 onwards;but it mustbe addedat once
that the chronologyof the Marcomanrucwars has not so far beensatisfactorily
established.
The first attack was made by the Lombards and Obii, 6,000 of whom at-
tackedthe sectionof the limes aroundBrigetio. However,two auxiliary units of
the Upper Pannonianarmy, one of which was stationedin Arrabona,succeeded
in repellingthemat once.I4 Dio associatesthis attackwith a largeembassywhich
petitioned Iallius Bassus,governor of Pannonia Superior, for peace.IS The
leader of this mission was Ballomarius, king of the Marcomanni,who led as
many as ten ambassadors, one from eachpeople.If this mission was connected
with the attackby the LombardsandObii, then onecannothelp seeingin it a last
attemptby the client-princesto procureRomanarmedaid. If other barbarians
wereresponsiblefor the invasion,why shouldit havebeenthe Marcomanniwho
askedfor peace?Aggressionby their northerly neighbours,however, could
well be usedas an argumentby them. It is very striking that elevenpeoplessent
emissariesto Carnuntumat the sametime andfor the samepurpose,thoughthey
hadnot previouslyundertakenjoint action; nor is anythingrecordedof a frontal
attack by elevenpeoplesduring the governorshipof Iallius Bassus.
It is, therefore,very likely that the first phaseof the war passedas a kind of
'dr6le de guerre' with Rome's allied neighboursjust beyond the frontiers of
Pannoniastill trying either to gain Rome's armed help or to achieve their
admissioninto the empire. Finally they were left no alternativebut to threaten
the prospectof a war if they werenot admittedinto the empire,nisi reciperentur.I6
This expressioncan be interpretedin two ways: admissioninto the empire by
assigningthem land within a province, or admissionby acceptingtheir terri-
tory as a province.It is impossibleto guesswhy Marcuscould not bring himself,
under suchcircumstances,to removethe impendingdangereither by meansof
armedhelp or by the receptio. Perhapshis experiencesover the years with far-
186
from-stable barbarian dynasties decided him against a military alliance; or
perhapsthe repulseof the Lombard attack: gave rise to the illusion that the
danger was less imminent than it was representedto be by the allied bar-
barians.
The government'shesitationresultedin Rome suddenlyseeingherselfcon-
fronted with a great barbariancoalition,17capablefrom about 167 of bringing
abouta seriesof catastrophicdefeats.What was presumablythe first shock was
also the most severe.TheMarcomanniand Quadi overranthe Pannonianlimes,
penetratednorthernItaly via the Alpine passes,and, after besiegingAquileia,
plunderedthe small town of Opitergium.IB The repercussionsof this invasion
were great,sincefor more than 200 yearsItaly had beenimmunefrom attackby
foreign foes.
The dateof this invasionis controversial,and the chronologyof the incidents
in the first yearsof the war is also uncertain.The year 168 is oneof the few fixed
points, when Marcus and Verusestablishedtheir headquartersaroundAquileia,
with what was presumablya considerablemilitary force.I9 This was the year,
too, that the plague reachedthe westernprovinces and Italy, and decimated
their populationas well as that of the barbarians.In 168 the emperorstook the
most necessarymeasuresfor the securingof Italy;20 and in this year too they
probablyestablishedthe specialcommandin the approachesof the Julian Alps
under the title praetenturaItaliae et Alpium21 and startedto clear the Danubian
provinces of the barbarianswho had invaded them.22 However, the plague
forced the emperorsto return to Romeat the end of the year; on the way, early
in 169, Verus suddenly died. The logic of the-admittedly little known-
eventssupportsthe dating of the siege of Aquileia and the sacking of Opiter-
gium beforethe emperors'stay at Aquileia ratherthanafter it. The year 167 was
a critical year in Daciatoo-asis proved aboveall by the wax tabletsfrom the
gold mines.23 The clearing of the barbariansfrom Pannoniaand Dacia was
directedby Marcusand Verus in person;thereare someindicationsthat Verus
was active in Dacia and Moesia.24 It was presumablyalso in 168 that legio V
Macedonicawas moved to the fortress of Potaissain Dacia.
It is unlikely that Marcus stayedlong in Rome. In 169 battles were fought
with the Germansand Sarmatianson all fronts; details howeverare unknown,
except that M. Claudius Fronto, an experiencedgeneralof the Parthianwar,
who was entrustedwith the joint governorshipof Dacia and Upper Moesia,
fought successfullyagainstthe Germansand Sarmatians(Iazyges)perhapseven
beforethe deathof Verus, but died in battlein 170 at the latest.25 The combina-
tion of UpperMoesiaandDaciain onecommandallows us to concludenot only
that the Romans had to deal with a frontal attack by the Iazyges from the
187
Crises on the Danube: the rise of the Ilfyrican soldiery
oppositelowlands but that the two reservelegions,IIII Flavia and VII Claudia,
had to be employedas well. The fact that the legions of UpperMoesiacould be
usedelsewhereshouldindicatethat this provincewas not in immediatedanger.
Moreover, at the latest after the death of Claudius Fronto the army of the
provincewas reducedby onelegion: this is shownby the fact that for sometime
the governorsof this province were of praetorianinsteadof consularrank.26
PannoniaInferior, on the other hand,with onelegion-wedo not know which
-addedto her army, was at that time possibly controlled by consulargovern-
ors.27 Of courseone cannot say that Upper Moesia was sparedaltogether:at
Singidunum a woman interfecta ab hoste is mentioned28 on a tombstoneand
similar casesare mentioned on several inscriptions in Pannoniaand Dacia;
while the hinterland of Singidunum in particular was heavily occupied by
auxiliary troops underMarcus for the purposeof protectingthe newly opened
mines of Mount Kosmaj.29
The focal point of the war was in Pannonia,and there it remained;Marcus
himself was presentin this province. In 171 the praetorianprefect Macrinius
Vindex was killed. 30 This loss is sufficiently indicative of the critical situation.
Neverthelessthe defeatseemsto have beenthe turning-point of the war. From
172 onwardsan offensivewas startedunderthe leadershipof ClaudiusPompei-
anus, Quintilius Maximus, Quintilius Condianus,and last but not least, Per-
tinax; Marcuswas continuallypresent,31andthe expeditionaryforce was formed
of vexillations of many legions. Not much is known about theseyears either.
Marcus stayedin Carnuntumfor two years and afterwardswent for a time to
Sirmium.32 He wrote part of his Meditationsin Pannonia:twice he mentionsat
the endof a book the placewherehe wrote it: one of thesewas Carnuntum,the
other the river Granua(Garam,Hron) in the country of the Quadi.
The sequenceof the expeditionswhich startedin 172 leads us to infer the
existenceof a well-consideredplan, which was to concentratethe army on one
region and severalneighbouringpeoples;the first offensivewas directedagainst
the Quadi and their neighbours(e.g. the Cotini), who formed the centreof the
coalition. The series of reliefs on the column of Marcus in Rome appearsto
startwith this offensive.It can bededucedfrom the Miracle of the Rain, which is
often mentionedin early Christian literature, that it again proved difficult for
the Romansto gain a victory. The Miracle of the Rain was not, in fact, experi-
encedby the army of Marcusbut by that of Pertinax, presumablyin the country
of the Cotini.33 The Quadi and some of their neighbourswere defeatedin the
end, and Marcus madeconditions devisedwith regardto the fact that the war
was not yet finished: the Quadi had to restoretheir captives(supposedlythirty
thousand),the cattle they had driven away and also, of course,the deserters.
188
They had to recognizea vassal king namedFurtius installed by Marcus, and
they were deniedaccessto the market-placesalong the limes; this last condition
was designedto enableRome to hold off the still undefeatedMarcomanniand
Iazyges. But the most importantand significant fact is the receptioninto the
empire by Marcus of some groups of Quadi and their settlementin several
Danubianprovincesandin Italy.34 The fact that hostiletribes weresettledin the
theatreof war in Pannonia,Moesia and Dacia, where they could easily make
commoncausewith the enemy,throws light on the characterof the whole war:
the settlementof barbariansin a threatenedprovince was permittedonly when
their pacification could be ensuredby the very fact of their admission.Similar
tendenciesare perceptiblein the caseof the Vandals who were admittedinto
Dacia. They causedthe governor considerableproblems,presumablybecause
they had only recently becomeneighboursof the empire.3 5 But of the Quadi
who becamemembersof the empire nothing further is heard.
Investitureof a vassalking was not the solution to the problemof the Quadi.
When Marcus undertook his expedition against the Marcomanniin 173, the
Quadi rebelledand dethronedFurtius, their king by the graceof Rome. Ario-
gaesus,proposedby the Quadi as their king, wantedto negotiatewith Marcus,
but, sincethe successof the Marcomannicexpeditionwas not at issue,the em-
peror did not enterinto negotiations.36 The conditionsof peacewhich the Mar-
comannihadto fulfil werenot different from thoseput to the Quadithe previous
year, exceptthat, as an additional measure,they had to vacatea broad strip of
land along the Danube.3 7 Operationsagainstthe neighboursof the Marcomanni
were also crowned with success.At the presenttime we only know that the
Naristae,a small tribe betweenthe Marcomanniand the Quadi, were defeated
and their king Valao killed by the hand of an auxiliary prefect.38 It can be as-
sumed that similar operations took place against other neighbours of the
Marcomanni, as well as those against the neighboursof the Quadi in 172.39
The measuresthat Marcustook for the pacificationof the DanubianGermans
were, to judgefrom what hasalreadybeensaid, of varied character.He usedthe
old device of investitureof subjectrulers, but combinedit with the transferof
groups of barbariansinto the empire, a method last used during the Julio-
Claudian period. The evacuationof a zone along the border, and thus the
establishmentof a no-man's-landon the approachesto the limes, was presumably
a temporarymeasure,dictatedby the critical situations,as was the prohibition
of accessto market-placesalong the frontier. It is not quite clear whether the
introductionof a stricter control over individualpeoples,implementedby army
officers, should be countedamongthe emergencymeasures.It is in connection
with the insurrectionof the Quadi during the Marcomannicexpeditionthat we
189
Crises on the Danube: the rise oj the Il{yrican soldiery
first hear of the Cotini ill-treating a Romanofficer; he was the celebratedTar-
runtenus Paternus,who was installed as their commander,apparentlywith
sometroopsat his disposal,after the conclusionof peacein 172.40In the second
phaseof the war every tribe probably receiveda Roman unit with an officer
chargedwith all aspectsof administration.Mter the Cotini had expelled Tar-
runtenusPaternus,they were defeatedand settledin Pannoniaeitherat onceor a
little later.41 At about the sametime the Naristaealso cameto Pannonia.42
The end of the year 173 saw only the Iazygesleft undefeated.The fact that
Marcus intendedto leave Pannoniaat the end of the year showsthat they did
not appeartoo dangerousnor their problemtoo pressing.But after the Iazyges
hadinvadedPannoniaacrossthe frozen Danubeduring the winter of 173-4,43
Marcus transferredhis headquartersto Sirmium and beganhis third expedition
from Pannonia.At the sametime the emergencymeasuresagainstthe Marco-
manniwere abandonedor mitigated:they were allowedto visit Romanmarkets
in certainplacesand on certaindays,and the evacuatedregion along the border
was narrowedby half.44 The Iazyges, on the other hand, had becomeall the
more difficult to handlesince their northerly neighbours,the Quadi, had given
them help.45 The king of the Quadi, Ariogaesus,who had also presumably
negotiatedwith Marcus in the meantime,apparentlyintendedto emphasizehis
demandsby supportingthe Iazyges.Marcus succeeded-we do not know how
or when-in quieteningthe Quadi, chiefly, perhaps,by managingto take Ario-
gaesusprisoner.46 The date of the victory over the Iazygeswas probably 175.
Their king, Banadaspus,had wished to negotiatewith the Romansafter their
first victory at the beginning of 174,47 but Marcus was not yet satisfied with
Roman successand refused the offer of submission,whereuponBanadaspus
was taken prisonerby his own subjects.Marcus enterednegotiationswith the
other king, Zanticus,only later, after a victory of greaterimportancehad been
won. The conditions were severe on this occasiontoo: the Iazyges had to
evacuatea zone along the border twice as wide as that evacuatedby the Mar-
comanni,andthey hadto providea contingentof 8,000men.Allegedly theygave
back 100,000 Roman captives.48 According to CassiusDio, Marcus intended
further measuresagainst them but was hindered by the rebellion of A vidius
Cassius.49 Thus the end of the great period of offensivesfalls in the year 175.
The secondphaseof the great war startedin 177. Marcus himself cannot
haveconsideredthe severeconditionsof peaceimposedin 173/5 to be a lasting
solution, but it was the barbarianswho beganhostilities. How Marcusintended
to solve the problemof the Danubianbarbarians,we do not know, particularly
as the new situation after 177 must have renderednew plans necessary.
The chronology of the years 177-80 is likewise not quite clear, but it has
19°
beenmore satisfactorilyestablishedthan that of the first phaseof the war. In
177 the barbariansinvadedPannonia;it remainsuncertainwhetherthe Germans
acted alone or whether the Iazyges joined them. Although Marcus soon re-
ceived an acclamationas imperator, the praetorianprefect BassaeusRufus was
killed this year50 and the othergeneralswere unableto obtainfurther successes.
In the summerof 178 the emperors-Marcusand Commodus-hadto appear
in personat the theatreof war.5I In the middle of 179 the Iazygesandthe Buri, a
tribe in the north-easternCarpathians,were defeatedand peacewas concluded
with the Iazyges. The conditions were that they had to stay away from the
islandsin the Danubeandthat they were not allowed to keepboatson the river.
On the other handthey were given certain concessions,aboveall permissionto
communicatethrough Dacia under the supervisionof the governorwith their
kindred on the lower Danube,the Roxolani, and to visit Romanmarket-places
on the bank of the Danubeon appointeddays.52
In the secondhalf of 179 Marcus formally occupiedthe country of the Mar-
comanniand Quadi and did not evacuatehis troops in the winter of 179/80.53
According to CassiusDio, the strengthof the Romanforces of occupationwas
40,000 men; if this numberis reliable, the army equalledin size the provincial
army of the two Pannonias.It seemsthat the whole expeditionaryforce win-
teredin barbarianterritory. Whetherin the last resortthis occupationwas meant
to lead to the formation of provincesout of the countriesof the Marcomanni
and the Quadi it is not possibleto deducefrom the eventsof the spring of 180,
when Marcus preventedthe Quadi, who were dissatisfied with the Roman
occupation-orrather, with the behaviourof the soldiers-fromemigratingto
the north.54 On the onehandit shouldbe possibleto concludefrom this that the
Quadi, after bitter experiencessuch as can be found daily during any occupa-
tion, wereno longerinterestedin becomingprovincials,whereasMarcuswanted
to pacify them as future subjects;on the other hand, CassiusDio quotesthis
episodein particularto prove that Marcusmerelyintendedto punishthe Quadi,
not to conquertheir land. He may be right. In the sameyear the governor of
Dacia, Vettius Sabinianus,causedthe emigration of the free Dacians to be
preventedand assignedland to them in the province of Dacia.5 5
After the deathof Marcus on 17 March 18o at Vindobona Commoduscon-
tinued the war. In August he gained a victory,56 which, as conclusionof the
expeditio Germanica secunda,57gave sufficient reasonfor the making of peace.
Commodusleft Pannoniafor Rome in the autumn of 18o. The conditions of
peace58 contained onlyone novelty, permanentcontrol of the internal political
life of the Marcomanni and Quadi. This control was exercisedby Roman
centurions,who probably officiated as praepositigentis.59 How long this control
191
lasted after 180 is unknown, but it was probably abolishedquite soon, and
certainlybeforethe deathof Commodus.It was basedwholly on the intimidation
of the defeatedpeople,and could not be sustainedwithout troops of occupation
stationedpermanentlyin barbarianterritory.
The mutilation of our historical sourceshardly permits a satisfactoryrecon-
struction of the new situation. CassiusDio's fragmentscontain only episode-
like summariesof the negotiations,missionsandpeace-treaties; yet differencesin
the emperor'streatmentof Germansand Sarmatianscan be observed,and we
may dare to draw severalgeneralconclusionsfrom them. The peoplewho re-
ceivedthe strictesttreatmentduring the first phaseof the war were the Iazyges,
against whom Marcus allegedly intended further punitive measures;they
neverthelesswere the easiestto quietenduring the secondphaseof the war, and
Marcusevenallowedsomeof their old demandsto be granted.TheMarcomanni
and Quadi, on the other hand, werenot as easily defeatedalthoughMarcus had
soon-perhapsas early as 174 or 175-alleviatedthe severeconditions of the
peaceof 172-3. Unlike the conditions imposed on the Iazyges,the measures
taken againstthe Germansin 179/80 were severein the extreme. Considering
that during the decadefollowing the Marcomannicwar the principal enemies
in the Carpathianbasin were neither the Suebi nor the Sarmatiansbut the
Vandalswho had madetheir appearance thereonly recently,it is very likely that
the specialmeasurestakenagainstthe Suebioriginatedin the recognitionof the
delicately balancedsituation among the free Germans-whichwas apparently
the result of the newly arrived tribes exerting pressureupon the Germanic
neighboursof the Danubianprovinces.It is, presumably,not the questionhere
of some irreconcilable German-Romanantagonism, becausein that case
Marcus would not have admitted some groups of Germansand their small
neighbouring tribes (Cotini, Naristae) into the empire, and settled them in
Pannonia, Moesia and Dacia. The Marcomanni, Quadi and several other
peopleswere so hard-pressedby the appearanceof the Vandals,Lombardsand
other Germanictribes that their only choice lay betweensubmissionto Rome
and submissionto other Germans. The bitter experiencesarising from the
'mutual armed aid' stipulatedby treaty were presumablythe reasonfor their
not readily resumingthe old client-relationshipwith the Romans.The danger
therefore existed that they would ally themselveswith their new Germanic
neighbours.Romansand Germanic tribes thus found themselvesconfronted
with the same alternative: alliance or war, with equally doubtful result. The
transferof severalgroupsfrom the left bank of the Danubeinto the empirewas
merely a palliative, not a solution of the problem.Marcus was presumablywell
aware of that, but he also presumablyappreciatedthat the incorporation of
192
Marcomanniain the empirewould createnew internal problemsand newneigh-
bours on the border.
The fact that the Iazygesallowed themselvesto be satisfiedwith the appoint-
ment of market days and permissionto associatewith the Roxolani via Dacia
indicatesthat their problemswereof a quite different nature.Nothingis reported
of a resettlementof groups of Iazygeswithin the empire; their commitmentin
175 to raise a considerablemountedcontingentpresumablymerely servedthe
purpose of weakening the military power of this nation of horsemen.The
Iazygespresumablymerely wantedto removethe disadvantagescausedby the
establishmentof Dacia; in other words, to establisha free communicationwith
the Sarmatianson the lower Danube.The control of trade was possiblyitself a
gain for the Iazyges.For conceivablyit was only from this time onwardsthat
they obtainedleaveto visit Romanmarketson the Danube.It is difficult other-
wise to explain why all the samianwarefound in the Hungarianplain datesfrom
the time of Marcus and the Severi, whereasin the northern half of the Car-
pathianbasin,which was inhabitedby Germansand Daco-Celts,a considerable
quantity of earlier samianhas beenfound.60
I considerit probablethat the solution which Commodusadoptedoriginated
with his father. Fulfilment of the plan to establishtwo new provinces,Marco-
manniaand Sarmatia,cannothave beenconsideredseriously afterall that had
happenedbetween167 and 180. It is significant that CassiusDio, the source
nearestin time to the events, who as the young son of a senatormust have
known personallya number of membersof Marcus' supremecommand,de-
pends on surmises.61 The other sourcesmerely repeatthe commonplacethat
Marcus would have annexedthe provinces to the empire, had his death not
preventedhim from doing SO.62 After the interpretationof eventsgiven above
we may be permittedto doubt whetherthe establishmentof Marcomanniaand
Sarmatiacould still havebeenseriouslyconsideredin 180. Dio corroboratesthis
in an indirect mannernot only by what he says about the preventionof the
emigrationof the Quadi but also in his descriptionof the conditions of peace
put to the Iazyges,which constituteda final solution, but shouldon no account
be regardedas preparationfor provincialization.On the whole, foreign policy,
after Marcus,continuedto treadthe old pathsof client-treaties,andit seemsthat
it proved possibleto embracethe new peoplesof the Carpathianbasin in this
policy basedon client-relationships.In the peace-treatyit was also stipulated
that the Germanswere not to wage wars with their neighbours,amongwhom
the Vandals are mentioned.63 From this we may conclude that Marcus and
Commodus succeededin creating a security system that included the new
neighboursas well.
193
If the Marcomannic wars changedthe foreign political situation on the
Danubeonly slightly, their effect on domesticpolicy was all the greater.First,
the Danubian provinces andin particular Pannonia,which had suffered the
heaviestlosses,had to be rebuilt and economicallyrestored;and, second,it has
becomemanifestthat the army of the Danube,representingabout one-thirdof
the empire'smilitary forces, and at the sametime the nearestto Italy of all the
.imperial armies, constituteda hitherto unesteemedbut potentially important
factor in the power politics of the empire. Reconstructionand the awakening
of self-confidencein the Danubelands thus went hand in hand. In this, by
degrees,the leading role was necessarilytaken by Pannonia.The Pannonian
provinces containedfour legions in the neighbourhoodof Italy; and this in-
volved a concentrationof economic support. With the economic and social
consequences of this favourable situation we shall deal in more detail in the
next chapter.
The lossessufferedby Pannoniawere heavyand various. We learn of whole-
saleremovalsof the provincial populationinto Germanor Sarmatiancaptivity.
(Allegedly the Germansgavebackthirty or fifty thousand,the Iazygesa hundred
thousandprisoners.)64 Cattle are reportedto havebeendriven away and booty
of all kinds carried off. The numerousfinds of coin-hoardsfrom the years of
war, and amongstthem such as were hidden during the secondphaseof the
war, are significant as well. Excavationsin civil settlementsand in forts have
revealedtime and againthe existenceof burnt layers of second-centurydate; it
has even becomea risky commonplacein archaeologyto relate all layers of
destruction and traces of burning to the Marcomannic wars. A thorough
critical sifting of thesetraceswould presumablyshow that more damagewas
donein thecivilian settlements thanat military sites,wheretotaldestructionwasan
exception.The greatestdevastationswere presumablyeffectedduring the years
167-71 and 177-8, whereasduring the years of the offensive, when Marcus
himself stayedin the theatreof war, the situation was comparativelynormal.65
No new period of construction,distinctly different from that previous to the
war, can be tracedin the limes-forts. A slightly protruding corner-towerwith
semicircular plan (Fig. 18, p. 108) and several newly added interval-towers
cannotso far be datedaccurately;the excavatorsattributethem to the reignS:of
Commodusor Caracalla.66 Of much greater importance were certain other
changesin the defenceof the frontier which had alreadybeenmadein the years
around 180.
In the first placethe transferof mountedarchersfrom Syria andAfrica to the
Danubemust be emphasized.Already in the first phaseof the war a strongunit
from the African legio III Augustawas employedagainstthe barbariansof the
194
Crises on the Danube.: the rise oj the Ilfyrican soldiery
Danube, the reasonpresumablybeing that the soldiers of this very distant
legion were used to the tactics of light cavalry. This African vexillation re-
mainedin Pannoniauntil 180 and part of it was not sentbackto Africa but was
incorporatedin legio II Adiutrix, 67 presumablyfor the purposeof training the
soldiersof the Pannonianlegion in the tacticsof light cavalry. UnderMarcusor
Gommodusseveral cohortesMaurorum were brought to the Danubefrontier
of Pannonia.68 The cohors I milliaria Hemesenorumequitatacivium Roman-
orum was transferredat the sametime to the front facing the Iazyges.69 It was
stationedin the fort of Intercisa,wherea small but importantSyrian settlement
developed.During the Severanperiod the cohort was further reinforcedfrom
Syria.
For similar changeson the Danubefrontier of Upper Moesiawe haveno evi-
dence,but from the interior of this province,whereMarcushadstationedmany
of the cohortesAureliae raisedby him, we havefurther information. The mines
at Mount Kosmaj, southof Singidunumand thereforequite nearthe Danubian
frontier, were presumablyopenedunder Marcus and organizedas a territorium
metalli. For the protection of these apparentlyvery important mines, several
auxiliary regimentswere transferredfrom the Danubianlimes in Upper Pan-
nonia;70thesewere soonreplacedby the cohortesAureliae novae7Iwhosefort
in Stojnik has alreadybeenexploredto some extent.72The fort's hospital was
built in 179,73but the stationingof auxiliary troopsat Mount Kosmaj is attested
almostten yearsearlier. Sincethe troopsfirst transferredto this placewere taken
from the army of Upper Pannonia,whereafter the outbreakof the greatwar all
provincial troopswould be needed,we must datethe organizationof the mines
and the transfer of troops from Upper Pannoniato the first years of Marcus'
reign; and it follows that the military occupationof the mines should be attri-
buted to the disturbanceof the internal peaceof the provinceby latrones; there
was a dangerof the metallarii making commoncausewith the latrones.
Even during the Marcomannic wars the latrones gave trouble to Marcus.
Specialcommandoshadto be sentagainstthem.Thesemayalsohavehadto solve
the problem of incorporating the latrones in the expeditionaryforce for the
Marcomannicwar. For the latrones Dardaniae this is attestedin the Historia
Augusta,74but it would be erroneousto assumethat the two cohortesAureliae
Dardanorumrecruitedby Marcusconsistedof former latrones.CohorsII Aurelia
Dardanorumwas assignedto the fort of cohorsI ThracumSyriacain Timacum
minus, which lay right in a centreof the latrones.7s The fort of cohorsI Aurelia
Dardanorumis unknown;it must be looked for somewherein the valley of the
Morava north of Naissusor on the northern slope of the Scardusmons, where
attacksby latronesarealsoknown to haveoccurred.76Both cohortesDardanorum
I95
were recruited from among the Romanizedpopulation of southern Moesia
(Dardania).Marcus' military measuresin Moesia Superior thus, for the most
part, concernedthe internal situationin the centreof the Balkan peninsula;the
decisivefactor was probablythe realizationthat the diagonalroad leadingfrom
Singidunumto Byzantium could becomeone of the most important military
roads of the empire. The towns situated on this road-Naissus,Horreum
Margi, Margum-wereprobably,without exception,promotedmunicipia under
Marcus. The eventsof the third centurywere fully to justify Marcus'foresight.
Another changein the defenceof the frontier is traceablethrough inscrip-
tions from PannoniaInferior. A numberof similar building-inscriptionsof the
year 185 attestnew buildings on the bank of the Danubebelow Aquincum.77
The inscriptions,which are uncommonlylong and complicatedin their formu-
lation for this period,reportthat Commodushadfortified the wholebankof the
Danube with burgi and praesidia; the burgi were built from scratch, and the
praesidia erectedto watch places most suitable for the secretcrossingsof the
latrunculi (Fig. 35). The word latrunculi hascausedcontroversyin that it hasbeen
taken to signify smugglers,rebellious subjectsof the empire, or simply bar-
barians,i.e. Iazyges.The very generalderogatorydiminutive (latrunculi) was not
intended to provide an accurateidentification of those against whom these
IMPCAES MAVRCOMMODVS ANTONINVS

AVG PIVS SARMGERBRPONBMAXTRIBPOT
VI IMP I COS P P RIPAM OMNEM BVRC S
ASOLOEXTRVCTIS ITEMPRAESIDIS PERLO
CAOPPORTVNAADCLANDESTIONOS LATRVNCV
OSITIS LATRVNCV
LORVMTRANSITVS OPP
PER L CORNELIVM FELICEM
PLOTIANVM LEG PRPR
Figure
35 Jf Inscription
Inscription of Commodusm
of Commodus entioning
mentioning latrunculi from
latrunculi from Intercisa
Intercisa
196
Crises on the Danube: the rise oj the Illyrican soldiery
praesidia were built; it included all who intendedto crossthe Danubeillegally.
Commodus'measures,attestedby theseinscriptions,were thus concernednot
with the betterprotectionof the frontier againstbarbarianattacks,but with the
supervisionand regulation of traffic acrossthe border. After what has been
saidthe necessityof thesemeasurescaneasilybe understood,comingas they did
a few years after the conclusionof peace with the Iazyges. Marcus had per-
mitted them to visit Romanmarketson stateddaysand had openedthe frontier
of Dacia to them under the control of its governor; thus under pretenceof
peacefulandlegal traffic acrossthe frontier small groupsof barbarianswere able
to raid the province,and they presumablyalso tried to tradeillegally outsidethe
appointedplacesand days.It was difficulties of just this kind which presumably
necessitated a punitive campaignagainstthe Sarmatians.The dateof this expedi-
tion is controversial;the Historia Augusta,78however,connectsit with an event
which is very characteristicof the alteredsituation in Pannonia,and which has
left tracesin the burgus-inscriptionsof Commodus.
It seemsthat only someof theseburgus-inscriptionswere actually attachedto
the newly built burgi; otherswere left in a stonemason's workshopat Intercisa,
and were later usedas slabsfor other purposes.The reasonis possiblythat the
governorof Lower Pannonia,CorneliusFelix Plotianus,hadfallen into disgrace
in the meantime:his namewas erasedon those burgus-inscriptionsthat had al-
readybeenfinished. This coincidesin time with the overthrowof the praetorian
prefect Tigidius Perennis,whose sons were at that time attachedas young
officers to the army of Pannonia.79 According to the Historia Augusta,Perennis
intendedto credit his sons with the successagainstthe Sarmatians;and their
presencein Pannoniawas connectedwith his further plans. He had evidently
realizedthat in the eventof a coup d'itat the four legions of the Pannonianarmy
were closestat hand for interventionin Italy. A slightly obscurephrasein the
Historia Augusta(Pannoniaequoquecompositae)80 allows us to supposethat the fall
of Perennisinvolved reprisalsin Pannoniaon a considerablescale,and we may
deducediscontentamong the Pannonianarmy againstthe rule of Commodus.
One Germancampaignof Commodus,his 'third GermanExpedition'cannot
be dated.81 It was presumablypunitive, like that againstthe Sarmatians.The
client-relationshipsestablishedby Marcus not only outlasted Commodusbut
presumablysurvivedin the main until the middle of the third century. The de-
partureof the Pannonianlegionsfor Italy under SeptimiusSeverusin the year
193 did not tempt the barbariansto invade Pannonia;accordingto Herodian,
Severusnegotiatedwith the tribes acrossthe frontier before his departure,and
inducedthem to keep quiet, presumablyby the samemeansas were usedunder
TampiusFlavianusin A.D. 69.82 Thereafterthe main centreof unrestlay on the
197
Criseson the Danube: the rise oj the Illyrican soldiery
Dacian limes, where the Vandals and several Dacian groups, above all the
Carpi, were the enemies.This unrestendangeredthe Danubianborderof Pan-
nonia only seldom and merely indirectly. Our sourcesfor Severantimes are,
however,so mutilatedand so unreliablethat researchhashad to rely to a large
extenton coin-hoards,votive inscriptionsand other indirect evidence.The re-
constructionof eventsthus obtained,however,hasnot only yielded too large a
numberof warson the Danubianborderbut hasbeenthoughtto revealbarbarian
invasionsdeepinto the provinceas well y andthis is difficult to reconcilewith the
mainly peacefulconditions of the Severanperiod. FurthermorCywe must re-
memberthatthe Pannonianlegions,or at leastvexillationsfrom them, took part
in almostall domesticandforeign wars underSeptimiusSevernsand Caracalla;
they formed, so to speak,the elite troops of the emperor.The reservelegion,
illI Flavi~ had to take the place of legio II Adiutrix at least once under Sep-
timius Severusin Aquincum.83 The Pannonianlegionswereemployedin Dada
as well againstthe barbarians,and this could not havehappenedif Pannoniaor
Moesia Superiorhad beensubjectto pressureby barbariansacrossthe frontier
at the sametime.
The only barbarianinvasion into Pannoruaunder the Severi which is to a
certain extent reliably documentedtook place under Caracallain 212 or early
213, and was connectedwith Carpian- Vandal unreston the northernfringe of
Dacia. The invasionaffectedthe north-easterncorner of Pannonia,84and pos-
sibly wasthedirect causeof Caracalla'salterationof the boundarybetweenUpper
and Lower Pannoniathe next year. This line had run from the bend of the
Danubesouth-westwards, andthe result was that a barbarianattack,onceit had
crossedthe Danubefrontier of Upper Pannoniafrom the north, could reach
Lower Pannonianterritory immediately:theretroopsof the governorof Lower
Pannoruahad to be employedagainstit. In 214 the borderwasthereforeshifted
further westwards,approximatelyto the mouth of the river Arabo. As a result
Brigetio with legio I Adiutrix joined PannoniaInferior, and so both Pannonias
becameprovinceswith two legions,and the governorsof Lower Pannoniawere
henceforthconsularstoO.85 Reasonsof statewere also presumablybehind this
decision, since three-legionprovinceswere abolishedby the Severi as having
armies of excessivesize and thereforegovernorsof excessivepower.
The eventsof c. 213 throw a clear light on the continuedexistenceand un-
disturbedfunctioning of the systemof client-kingdomson the Danubefrontier.
According to CassiusDia, Caracallaboastedof havingcreatedhostility between
the Marcomanniand the Vandals.86 According to Herodian'shighly imagina-
tive account,Caracallatook pains duringhis visit to the Danubianprovincesto
poseasa friend of theGermans.87 Presumablyhe tried hardto isolatethe Danub-
198
Crises on the Danube,'the rise oj the Il!Jrican soldiery
ian Suebi (Marcomanni and Quadi) from the increasingly hostile Vandals;
with the Marcomannihe succeeded.The Quadi, on the other hand, who in-
vadedPannoniain 212 or 213, he could punishon the basisof the client-treaty
without any particular consequences:the Quadi had to surrendertheir king
Gaiobomarusand severalnoblesof the tribe to him, and theseCaracallasimply
orderedto be executed.88
The Severiwerethus successfulin maintainingthe systemof client-kingdoms
althoughpresumablyagainstincreasingdifficulties. The existenceof permanent
and well-paid interpreters,whose position on the governor's staff was pre-
sumablyfirst establishedunderthe Severi,showshow much attentionwas paid
to diplomatic relationsat that time.89 In provinceswherethe native population
spokea dialect not unlike that of the barbarianssuchinterpreterswerepresum-
ably not necessary,particularly sincea large proportionof the soldiers,part of
whoseduty it wasto servein the ofIicium consularis,werecapableof understanding
the languageof barbarianemissarieswithout difficulties. The fact that in Pan-
nonia interpreters,whoseofficial namewas interpresofIicii cos. salariarius legionis,
areattestedfor Lower Pannoniaonly, presumablyhasits explanationin the fact
that the governorof UpperPannoniahad Germansoldiersat his disposal.In the
ofIicium of the governor of Lower Pannoniainterpretersare attestedfor three
languages:Sarmatian,Germanand Dacian. The needto employ Dacian inter-
pretersas well arosefrom the Dacianizationof the Celtic and Germanicpeoples
on the northern fringe of the Lowlands which had begun long before. The
Cotini settled under Marcus in Pannonia around Mursa and Cibalae bore
Daco-Thracian namesfor the most part.90 Interpreterswould not have been
necessaryif knowledge of the Latin languagecould have been presupposed
among the barbarianprinces, and if it had not been necessaryto maintain
continual contact with the barbarians,whether of a political or economic
kind.
The basicallypeacefulsituationin Pannoniawas advantageous to the Severi,
the main supportfor whosepower was the Pannonianarmy. Of course,we do
not know how much moneythe treasuryhad to pay for this peaceon the Pan-
nonianfrontier. The emperorsmust probablyhave attachedspecialimportance
to the Pannonianlegions being always in a stateof preparedness, and for that
reasongrudgedno pains to createa modusvivendi with the Danubianfrontier
peoples.Caracalla,the 'friend of the Germans'in Herodian,playedhis partin this
aspectof the foreign policy of the Severi.We have alreadydrawn attention,in
connectionwith the overthrow of Perennis,to the ability of Pannoniain the
eventof a coup d'etat to playa major part becauseof its four legions,and even
more becauseof its nearnessto Italy. This fact had already becomeobvious
199
during the Year of the Four Emperors,but the internal situation of the empire
did not give opportunityfor anothercenturyfor the Pannonianarmy to exploit
the arcanumimperii. The Marcomannicwars were the first to demonstratehow
quickly Italy could be reachedfrom Pannonia.This was presumablyrealizedby
the Pannoniansoldiers,and, in his numerousadlocutionesduring the Marcoman-
nic war, the Emperor Marcus obviously did not conceal the fact. The key
position occupiedby Illyricum during the third century has already becomea
commonplace.9I
The attempt of Tigidius Perennisto stir up the Pannonianarmy against
Commoduswas not successful;but seven years later, when the governor of
PannoniaSuperior,L. SeptimiusSeverus,was proclaimedemperoron 9 April 92
in Carnuntum,Rome and the empire found themselvesconfrontedby an army
marchingwith astonishingspirit and discipline behindits candidate.Only in the
caseof legio X Geminais there a suspicionthat it did not take part in the pro-
clamation of the new emperor;93 but it is extremely unlikely that this legion
offered resistance,stationed as it was in the immediate neighbourhoodof
Carnuntum.At any rate, X Geminadid take part in the struggleagainstPescen-
nius Niger.94 After the troops of the Danubian provinces had marchedinto
Rome with Septimius Severusthe new praetorianguards were chosenfrom
amongsttheir men,95mainly from Pannonians,but to a lesserbut still consider-
able degreefrom Moesians,Thraciansand Dacians.The future recruitmentof
the praetoriancohorts,too, was securedfrom theseDanubianprovinces,service
in the guardsservingat the sametime as the first stepto a highermilitary career.
Legio II Parthica, stationednot far from Rome, also consistedof Danubian
soldiers. Besides the Pannonianlegions, the legions of the other Danubian
provincestoo were from the beginningsupportersof the regime of the Severi.
For this reasonthe collectivenameIllyricum beganto be usedimmediatelyafter
Septimius Severus'seizureof the throne. Gradually this name gained special
political meaning.As early as Herodianwe find the word Illyricum in useas an
inclusiveterm for the Danubianprovinces;96 its origin is not clear.In the second
centurythis term was usedfor the Danubianprovincesonly in the customsad-
ministration(publicumportorium Illyrici), and in Appian'sIllyrike the statementis
found that the word Illyricum has no linguistic-ethniccontent.97 It is therefore
quite possible that the name of the customs district was transferredto the
Danubianprovincesin general.Modern historiography,much too sensitiveto
national and racial differences,is inclined to attribute the part played by the
Danubianmilitary to a nationalself-assertionof the Illyrian race;but we shall see
that this Illyrian military consistedof Celts, Thracians,Illyrians, evenof Roman-
ized city-dwellers of mixed origin, and of other constituentsof the empirewho
200
migratedto Illyricum. From SeptimiusSeverusand his dynastyonwardstill the
middle of the third century,the representatives ofIllyricum werenot evenpeople
from the Danubebut mostly senatorsfrom other parts of the empire; people
who were Illyriciani by birth cameinto power only in the last third of the third
century.
The manythousandDanubiansoldiersservingin andaroundRomewereable
to exert a not inconsiderablepressureon politics in Rome. They behaved,in-
deed, as representativesof the interests of their homeland,expressingthese
interestsfrom time to time vigorously. Best known is the caseof the historian
CassiusDio, who in 229 was prevented bythe threatsof the praetoriansfrom
attendingin personthe inaugurationof his secondconsulship;their reasonwas
that as governor of Upper PannoniaDio had previously implementedun-
popular policies.98 Similar occurrencessoon producedan attitude of reserve
towardstheIllyriciani amongsenatorialcirclesandthe representatives of various
other provinces.This attitudehad a welcomepretextin the lack of humanitas,in
other words the rural roughnessand low level of education,of thesesoldiers.
In the previouscenturyFronto had contrastedthe clumsinessof the Pannonians
with the sophisticationof the Syrians.9 9 Herodian attributed the seizure of
power by Septimius Severusto the stupidity of the Pannonians;IOO and the
urban Roman population, plebs as well as senators,watchedwith horror the
presenceof thesebarbarianswho did not evenunderstandLatin well. 101 During
the first yearsof the dynastyof the Severi,the fear of Illyricum was presumably
employedas an elementin power-politics; it was thus used, for instance,by
SeptimiusSeverusin 196 when, on settingout againstClodius Albinus, he com-
mitted his son Caracallato the care of the Danubearmy after having him pro-
claimed Caesarin Viminacium in the summerof that sameyear.I°2 The young
Caesarreceivedthe news of the victory over Albinus in Pannonia.I03In 202 the
emperors,coming from the Orient, travelled throughthe Danubianprovinces,
and by journeyingalong the limes took specialcareto visit the troopsof Moesia
and Pannoniain their forts.I°4 It is, moreover,not unlikely that the journey
from Carnuntumto Romewas arrangedin sucha manneras to enableSeverus
to be in Carnuntumon the dies imperii, 9 April, and to enterRomeon the day of
his state entry nine years before. By 202 the great days of the Illyriciani were
already over; PescenniusNiger and Clodius Albinus were defeated,the war
against the Parthiansfinished. The legions were in their fortressesagain: II
Adiutrix hadperhapsreturnedwith the emperorsfrom the East-andSeptimius
Severusperhapswantedto demonstratebefore the troops as well as before the
empire at large that Illyricum was to remain the most favoured part of the
empire.
201
Sovereignty,however,cannotbe basedon so limited a socialfoundation.The
interestsof the Illyriciani came more and more to be consideredas sectional
and thereare a few indicationsthat as early as Alexanderthe Pannoniantroops
were not quite satisfiedwith the government.IOS Betweenforeign and internal
politics there existeda connectionin that the Danubian troopsdid not have to
advocateconcentrationof the empire's resourcesin Illyricum as long as the
Danubianborder was quiet-andunderAlexander, especially,the peaceseems
to havebeenundisturbed.But a seriousproblemarisesfrom the fact that during
the following decadesthe Illyriciani wishedto makeuse of the resourcesof the
empireto protecttheir homeland.Sucha demandmay perhapshavebeenmade
for the first time during Alexander's Persian war, when the Rhenish and
Danubiantroops desiredto return home on the news of a war with the Ger-
mans.106
The outbreakof this war against the Germanson the Danubeis not well
authenticated,and it is by no means impossible that the German dangerin
Illyricum was a false rumour. Coin-hoardswith latestcoins of the years228 and
2. 3I respectivelysuggeston the contrarya Sarmatianattack,crossingthe Danube
border of Lower Pannonia.I07 It is not mentionedin written sources.Perhaps
this Sarmatianattack was the first beginning of the Sarmatianwar, fought by
MaximinusThraxfrom 2. 36 to 2. 38. During theseyearstheimperial residencewas
at Sirmium,108so usedfor the first time sinceMarcus had stayedtherein 174-5,
but destinedto be by no meansthe last.We mustdwell on this pointa little, since
Sirmium and its surroundingsplayedan increasinglyimportantpart in power-
politics and strategyfrom the middle of the third centuryonwards.We haveal-
readyseenthat muchearlierthe Scordisciowedtheir temporaryhegemonyto the
circumstancethat they were inpossession of the areaaroundSirmium, a country
lying at the focus of routes connectingItaly and the west with the Balkans
and the Orient (p. 10). After the Goths made their appearanceon the lower
Danube,and at a time when the Rhine and Danubelegionshad to be employed
repeatedlyagainstthe Parthiansandthe Persians,deploymentof troopsthrough
Pannoniaandthe Balkanpeninsulabecamenecessary.As we haveseen,this had
alreadybeenrecognizedby Marcus,andit was presumablynot by chancethat he
chose Sirmium as headquartersagainst the Sarmatians.In the third century
Sirmium gradually becameone of the great headquartersof the empire.I09
The results were twofold: on the one hand economicprosperityprevailedin
the area, but on the other a clique of Sirmians arosewho often succeededin
obtaining empire-wide power. All Pannonianemperorsoriginated from the
Sirmium region, and someemperorsalthoughdemonstrablyfrom otherIllyrian
provinces have been taken for Sirmians by some of our sources.IIO The first
202
Crises 011 the Dal1ube: the rise oj the Ilfyrical1 soldiery
emperorfrom the Sirmian region was Decius; the last was Valentinian, who
camefrom Cibalae.
It is not unequivocally establishedthat the enemiesin Maximinus Thrax's
Sarmatianwar were indeed the Sarmatiansof the Lowlands: Dacia, not Pan-
nonia seemsto havebeenthe provinceto be exposedto the heaviestpressureby
barbarianattacksuntil the middle of the third century, andthe attackson Pan-
noniathus often camefrom barbariansliving on the northernfringe of the Low-
lands and in the neighbourhoodof Dacia, as for instancethe Vandals and the
Carpi. Emperorsassumedthe title Sarmaticuswheneverthe ripa Sarmaticawas
attacked,irrespectiveof the defeatedpeoplesbeing really Sarmatiansor not.
More recentresearchhas endeavouredto fill the gapsleft in the direct sources
by certain hypotheses,which might enableus to reconstructthe strugglesthat
took place on the Danubelimes in the middle of the third century. It has been
assumedthat imperial surnamesgrantedto the troops (e.g. legio II Adiutrix
Philippiana) were distinctions for martial successes,and accordingly that bar-
barianshadto be fought on the PannonianDanubelimes, especiallyin the north-
eastpart of the province, in the reigns of Gordianus,Philippus, Trebonianus
Gallus, etc. If this hypothesisbasedon the imperial surnamesis correct, the
questionstill remainsto be answeredwhetherthe troops concernedwere given
their titles for wars in Pannonia,or becauseof their part in an expeditionout-
side Pannonia.The Pannonianand Moesian troops presumablyquite often
took part in the wars in Dacia and MoesiaInferior, althoughthis can rarely be
provedbecauseof the lack of suitableinformation in our sources.
Coin-hoardsfrom the time of TrebonianusGallus certainly allow the conclu-
sion that after 252 the Quadi becamerestlessand attackedthe limes around
Brigetio.l l I This conclusion is not unreasonable,for the reason that, to all
appearance, both Marcomanniand Quadi were amongthoseresponsiblefor the
greatcatastropheof the years25 8-60. But the first certainly attested struggles in
Pannoniaare those which occurred under Gallienus; these had vast conse-
quences.Although the situation in the provincesof Illyricum becameincreas-
ingly dangerousafter the forties of the third century,at first a situationof chronic
war developedonly in Dacia and on the lower Danube,where the Goths and
the peoplesunder pressurefrom them were an increasingdangerto the limes.
The representativesof Illyricum in the imperial court presumablydemanded
energeticmeasures,and when Philippus in 247 personallytook part in the diffi-
cult war against the Carpi, he had the opportunity to convince himself that
extraordinary measuresmust be taken to avert the impending danger. The
obvious stepwas to unite the military forces of Illyricum undera supremecom-
mandof their own-thusin effect following the exampleset in the years I 17-19
203
Crises on the Danube: the rise oj the I1fyrican soldiery
or 168-70, when Marcius Turbo and Claudius Fronto had been entrustedres-
pectively with the simultaneousgovernorship of several adjacent Danubian
provinces.Possiblyas early as his return march,but certainly at the beginningof
248, Philippus appointedTi. Claudius Marinus Pacatianusas commanderof all
Danubian troops or at least of the troops of PannoniaInferior and Moesia
Superior,II2 The constitutional standing of this special commission, which
henceforthwas given frequently, is not clear,II3 Probably the supremecom-
mander was not simultaneouslygovernor of all the provinces concerned,but
merely a military commanderwithout administrativeauthority in the civilian
sphere.His seat seemsto have been Sirmium; this, too, could be indicative of
the intention not to impair the governor'sposition,for Sirmium was neitherthe
residenceof a governornor a legionaryfortress.
It seemsthat the Illyrican soldiery had only waited for a supremecommander
of their own to proclaim him emperoras their representative,andit must not be
forgotten that they could count on the aid of the praetoriansand the other
Illyriciani in Rome,and thus had good reasonto believethat they would be able
to repeat the march of Septimius Severuson Rome. A devastatingattack on
Lower Moesiaby the Goths and many other peoplespreventedPacatianusfrom
turning against Philippus, and his troops soon felt induced to dismiss him,
presumablybecauseof his failure or his inability. A decisivefactor in their deci-
sion, however, was the dispatchof Decius to Illyricum: when the news of the
seizureof power by Pacatianusand other similar news reachedRome, and when
Philippus had let his intention to abdicatebe known, the senator C. Messius
TraianusDecius,a native of Sirmium, wassentto Illyricum againstPacatianus,II4
Decisive for Philippus' choice of Decius was presumablythe assumptionthat
as a Sirmian he would be welcome to the Illyriciani. Decius, who apparently
knew his countrymenbetter and was presumablyalso better informed of the
intentionsof the leadingIllyriciani, had warnedthe emperornot to entrusthim
with the supremecommandin Illyricum; but he hadin the endto set out against
Pacatianus.Pacatianuswas killed, probablyas a direct result of the news that the
Sirmian Decius had been sent out as new supremecommander,and Decius
managedto fight successfullyagainst the Goths. But his successesgave rise
amongstthe soldieryof the Illyriciani to the idea that Deciuswould be the right
man for the realization of their plans. Against his will Decius was proclaimed
emperorin the summerof 249 and was forced to marchagainstPhilippus.
The circumstancesof the proclamationof the first Pannonianto becomeem-
peror show that the Illyriciani were keen on finding a suitable representative
capableof mobilizing againstthe increasingdangerfrom the Goths not only the
forces of Illyricum but thoseof the whole empireas well. The chief problemto
204
be solvedby eachsupremecommanderwas victory over the Goths,the imperial
throne being merely a meansthereto. After the death in battle of Decius the
procedureremained the same. In the place of Decius, TrebonianusGallus,
governor of the most seriously threatenedprovince,MoesiaInferior, was pro-
claimed; but as soonas a generalcapableof dealing successfullywith the Goths
emergedin the personof Aemilius Aemilianus, supremecommanderin Illyri-
cum, the purple was given to him. Aemilianus was forced, as Decius had been
beforehim, to set out againstItaly, but he did not succeedin maintaininghim-
self there, and after having defeatedthe Emperior TrebonianusGallus he suc-
cumbedto Valerian.
After 25 3 there was a suddencalm as far as the Illyrican soldiery was con-
cerned. Since the governmentof Valerian and later of Gallienus was certainly
somewhatanti-Illyrican we mustnot ignore thepossibility thattheyhadremoved
the representativesof the Illyrican soldiery from the most important offices. It
is at any rate striking that after 253 no supremecommanderis attestedin Illy-
ricum. V/ e may deduceeither that none was appointedor that he successfully
resistedthe intentionsof the Illyriciani. Nothing is said about larger attacksby
barbariansat this time; there was thus no immediateinducementto proclaim an
emperorin ordC'r to fight them. This is confirmedby eventsfrom c. 256 onwards.
About this year or at the latestin 257 the youngerValerianus,son of Gallienus,
was appointedsupremecommanderof Illyricum; this shows clearly that the
situation on the borders had becomeaggravatedto such an extent that the
appointmentof a supremecommander-whileperhapsat the sametime placat-
ing the demandof the III yriciani-had becomeimperative. However, the com-
mandersentwas a youth, who althoughin the confidenceof the emperorcannot
be considereda general.
The chronologyof ever,ts between257 and 260 is far from clear or attestable
in sources. We do not ifltend to add to the vast number of chronological
hypotheses,nor do we wish to follow any of them. Everythingis as yet unfixed
and will remain so, until new finds give new information. Presumablyit is
better to enumeratethe individual facts without attempting to relate them
chronologically.
In 257 Gallienus took the titles DacicusMaximus;II 5 it follows that the pro-
vince of Dacia, continually overrun since Philippus, bore the brunt of this new
attack as well. Pannoniawas probably invadedsomewhatlater, but then all the
more heavily; the number of coin-hoards-overa dozen-datableto the years
258-60illustratesthe seriousness of the attack.The coin-hoardsare evenly distri-
butedover the whole areaof the province,II6and attacksby Suebiand Sarmati-
ans are alike attested.II7 The attacks also affected part of the Moesian border
205
Crises on the Danube: the rise of the I1!Jrzcan soldiery
along the Danube,wherecoin-hoardsare also known.lI8 Up to this time Upper
Moesiahadbeensparedheavybarbarianattacks;theprovinceowedits immunity
solely to its favourable geographicalsituation. There are, in fact, a few coin-
hoardsfrom UpperMoesia-someof them huge-whichwere concealedduring
the reignof the first emperorsof the 'Military Anarchy'.119But we believethat
theseshould rather be attributedto internal disturbances,in particular to con-
centrationsof Illyrian troopsunderthe leadershipof usurpers:the generalswho
were proclaimed emperorsby the troops of Illyricum mostly fought on the
lower Danubeand in Moesia Inferior, and from there they had to make their
way to Italy via UpperMoesia,andin particularalongthe Naissus-Singidunum-
Sirmium road.
In Pannonia'scritical situationpretendersproclaimedthemselvesoncemore.
Probably they were supremecommandersof the Illyrian army as Pacatianus,
Decius and Aemilianus had been. The first one was Ingenuus,12owho was per-
haps appointed to replace the younger Valerianus now deceasedor killed.
Characteristicallyenoughfor his own position and the part playedby Sirmium,
he was proclaimedemperorin this town. Ingenuuswas supportedby Pannonian
and Moesian troops, but he was soon defeatedby Aureolus near Mursa.I21
The secondusurperwas the senatorP. q ...) Regalianus,who was also pro-
claimedin Sirmium, and by Pannonianand Moesiantroops.IU We know more
about this man than we do about his predecessor;perhapswe can deducefrom
this fact a comparativelylonger reign. It is reportedthat he had fought success-
fully againstthe Sarmatians.Towardsthe end of his reign he permittedcoins to
be struck, which circulatedin the north-westernpart of Pannonia(Fig. 36) and
aboveall at Carnuntum.I23 Thesecoins were struck in greathasteand-forwant
of raw material-from older coins overstamped. The denarii of a senatorial
lady, Sulpicia Dryantilla, overstruckin the sameway, belong also to this group
of numismatic rarities. This Sulpicia Dryantilla accordingly was either the
mother or the wife of Regalianus.But it is curious that more examplesof the
coins of Dryantilla are known than of those of Re:salianus.From the distribu-
tion it could be deducedthat Regalianushad to withdraw to the north-western
cornerof Pannoniatowardsthe endof his reign, wherehe thenhadthe irregular
coins struck-possiblyfor mere propagandapurposes.
During theseperiodsof unrestGallienushad to go to Pannoniaat leastonce.
One episode,otherwisenot datable,can be attributedto theseyears by the fact
that it presupposesGallienus' presencein person: accordingto this story the
emperorbecameinfatuatedwith the daughterPipa of the Marcomannicking,
Attalus, and thereforetransferreda part of Pannoniato him.124 This surrender
or transferof a pieceof land of the empiremust, of course,be understoodin the
206
Crises on the Danube: the rise of the Il!Jrican soldiery
o 100 200 km
Danube:
Danube:
Danube:
Figure}6 Distribution of coins of Regalianusand Dryantilla
senseof a group of Marcomannibeing settledby Gallienus in Pannonia.And

since the Marcomannihad taken part in the war of 258-60 we may place the
settlementof somegroups of them in Pannoniawithin theseyears.
For the further events,possiblythe most significant fact was that Gallienus-
presumablythrough his generalAureolus-broughta contingentof probably
considerablestrength,consistingof vexillations of the British and Germanpro-
vincial armies, to Pannoniaand stationedthem at Sirmium.I25 They remained
in Sirmium till after the years of crisis 25 8-60: this is attestedby their taking
part in the battle againstMacrianusin the BalkansnearSerdica(bellum Serdicense)
whencethey returnedto Sirmium.I26 Furthermore,they could not be sent home
becauseof the Gallic empire. By their presencein Sirmium they held the Illyri-
ciani, whoseleading clique was domiciled there, in check, and at the sametime
they occupiedthe most important place in the Danuberegion.
Regalianuswas killed by his own soldiers,if this information in the Historia
Augustais to be trusted.Sucha murderwas in itself by no meansunusual,since
207
Crises on the Danube: the rise oj the Il!Jrican soldiery
the emperorswho were proclaimedby the Illyriciani were treatedas the tools of
a junta whosepolitical intentionswerenot decidedby personalpoints of view.I27
It is striking that the end of the usurpationsin Illyricum coincideswith the im-
portantmilitary and political reformsof Gallienus,which were directedtowards
the organizationof a mobile centralarmy. Possiblythe British andGermanvexil-
lations were one of the reasonswhich inducedGallienusto establishthis central
army. Gallienus was intelligent and wise enoughto have the Illyriciani repre-
sentedin the newly organizedarmy; at leastthe so-calledlegionarycoins of this
emperor,probablyreferring to the new central army, allow us to concludethat
the legio1tes Il!Jricianae too had provided detachments. I28 Under such circum-
stancesthe junta of the Illyriciani had no choice but to make sure themselves
that their interestswere understoodin this new concentrationof power, and
thereforethey gave up the tactics of local usurpations.The fall of Regalianus
possibly belongsto the period of theseevents.
During the remaining years of the reign of Gallienus the interior of the
Danubianprovincesremainedcalm, nor were thereany conflicts of importance
on the limes. Pannoniahad, however, suffered somuch during the crisis years
258-60that this temporaryquiet undera governmentfacedwith the most severe
difficulties both in foreign and in homepolicy did not makeany reconstruction
or consolidationpossible.Moreover, a further circumstance,not to be under-
ratedin importance,has to be consideredhere: for obviousreasonsthe govern-
ment of Gallienusdid not considerit a major duty to make useof the resources
of the empirefor the reconstructionof Pannonia:first, becausetheseresources
werenot sufficient anyway;second,becauseseveralotherprovinces,amongthem
Moesia Inferior and Dacia in particular, neededreconstructionjust as much;
and, third, becausethe experiencesof the last two decadeshad provided ample
reasonsfor not considering the demands of the Illyriciani. One important
measure,however,has to be mentioned:the establishmentof the mint in Siscia,
which was to becomeone of the principal mints of the empireand the most im-
portant one in the Danubian provinces.Thereby,the supply of our provinces
with inflationary coinagewas securedon the spot. The raw material was pro-
vided by the mines of Dalmatiaand Moesia; the reasonfor choosingSiscia for
the mint was presumably that this town, situated fairly near to Italy, had
good connectionswith the mines,I2 9 At the sametime the mint ofViminacium,
the only placeto producecolonial coinageof the areaof the middle Danube,was
closed.This mint ofViminacium, which had beenin productionfrom the time
of Gordian to the time of Gallienus, in addition to the bronze coinage of
Col(011ia) Vim(inacium)P(rovinciae) M(oesiaeS(uperioris)hadalso struck the bronze
coins of the Provincia Dacia.I30
208
Crises on the Damtbe: the rise oj the ]lfyrican soldiery
Thereis no need,I trust, to prove in detail that the appearanceof the Goths
on the lower Danube and of the Gepidae in the Carpathian basin created
an entirely new foreign situation which was soon to destroy the old client-
relationships(which had functioned well during the time of the Severi) on the
lower Danubeandin the easternfringe of Daciaas well as in the Carpathianbasin.
The frontal attack on Pannoniain the years 258-60 has quite a few featuresin
commonwith the first phaseof the Marcomannicwars (167-71),when all neigh-
bouring peoples possessinga client-relationshipwith Rome entered upon a
periodof unrest.And the reasonsfor this attackwere presumablythe same,too.
At any rate we must interpretthe settlementof Marcomanniin Pannoniain the
sameway as the settlementsby Marcusof Suebiand Cotini in Pannoniaafter the
first war: the old neighbouringpeople wished to obtain the receptio by force,
al}d this receptio was grantedto a few groups.This time the peoplesunderpres-
surewere no longer only the Suebi and Sarmatians,but the Vandalsand Carpi
as wen who were exposedto the pressureof the Goths and the Gepidae.After
260 hca-vy fighting betweenGepidae,Vandalsand Sarmatianstook place in the
Carpathianbasin, outsidethe provincesPannoniaand Dacia. The Gepidaewho
had penetratedinto the Carpathianbasin soon after the middle of the third
centuryI3I wishedto drive away the Vandals,who sincethe time of Marcushad
settledon the northernfringe of Dacia; whilst from the Southa new Sarmatian
tribe, the Roxolani, invadedthe Lowlands. The Roxolani, who were forced to
evadethe pressureof the Goths, migratedthrough Dacia to their kindred the
Iazyges,presumablyby the route that Marcushad permittedthem to use. There
are someindications that they had alreadyappearedon the border of Pannonia
c. 25 8-60.!32 The most seriouslyendangered areawas thereforeDacia,which had
the Carpi and Goths as neighbourson the east,the Gepidaeand Vandalson the
north and Vandals and Roxolani on the west. Under such circumstancesthe
evacuationof part of DaciaunderGallienuscannotbe passedover lightly. Under
Gallienus the two legions of Dacia, V Macedonicaand XIII Gemina, appear
with their officers and their administrativeranks in Poetovio,where they richly
furnished a Mithraeum with sculpturesand votive inscriptions.!33This shows
that after 260 the army of Pannoniahad beenincreasedby two legions, appar-
ently for the protection of Italy but at the expenseof Dacia, which after the
appearanceof the Goths and Gepidaehad becomea causefor war rather than a
strong-pointfor the protectionof the Danubefrontier. It is possiblethat Dacia
had already beenthe subjectof negotiationsbetweenGoths and Romans,and
as a result of thesehad beenpartially evacuatedalreadyat the beginningof the
fifties.I 34 This episodeis, however,little understoodas yet, and does not come
within the scopeof this work.
209
The relative calm on the Danubeafter 260 presumablyhas its causepartly in
the mutual strugglesof the Gepidae,Vandals and Sarmatiansand partly in the
preparationsof the Goths for the invasion of the Mediterraneanby sea. These
Gothic attacks on the coastal regions of Asia Minor, Thrace and Macedonia
sooneror later made it clear that the preservationof Dacia not only had no
strategicadvantagesfor the empire-inparticularsincethe Gothscould advance
on the heart of the empire-butthat it had becomea burden on the imperial
army. The final pmof was given by the great attack of the year 268, when the
Balkan provinceswere laid waste from the southby the Goths,who did not
evenspareAthens.They invadedMoesiafrom the southandadvancedalongthe
Thessalonica-Scupi-Ulpianum road as far as Naissus; it was only there that
Gallienus, who in the meantimehad arrived with the mountedcentral army,
could defeatthem. Soonafterwards,Gallienushadto turn againstAureolus,a.11d
he left the war againstthe Goths in the Balkans in the hands of his ge.t;ieral
Marcianus.I 35
Now camethe great momentof the Illyriciani, who, as protectoresin the en-
tourageof the emperorand in the supremecommandof the mountedcentral
army, had gradually got the upper hand during recentyears. The exclusionof
the senatorialorder from the highestposts of the stateand army by Gallienus
was a favourablecircumstancefor the Illyriciani. They soon succeededin get-
ting rid of Gallienusas well as of Aureolus and in proclair.aingtheir candidate,
the DardanianM. Aurelius Claudius,emperor.The leadingmen of the Illyrican
soldiery at that time were the praetorianprefectHeradianus,the later emperor
L. Domitius Aurelianus,and the EmperorClaudius)probably all of them born
in Moesia Superior.136 Whetheran expressionof the local interestsof the pro-
vince Moesia Superior,which had sufferedmuct especiallyfrom the Goths, is
j,
to be seenhere may be left undecided.At any :late Claudius soon went to the
sceneof the strugglein the Balkans,where he remaineduntil his deathat Sir-
mium in the spring of 270.
Claudiusdied from the plague,which was then breakingout as it had a hun-
dred years before. The highest representativesof the Illyriciani, assembledin
Sirmium, did not follow the choice of the Senatein Rome; they refused to
acceptClaudius'brother,Quintillus, as emperor,but proclaimedAurelianusthe
secondhighestman in their clique. This man, regardedby later tradition as an
'honoraryPannonian',a Sirmian,137was to be the man who restoredthe unity
of the empire. To him is due much of the reputationof the Illyriciani as the
'savioursof the empire',as satisoptimi rei publicae.I38 Among the regional juntas
the Illyriciani, indeed,appearto havebeenthe only oneto havea feeling for the
unity of the empire. Rearedin a strategicallyimportant part of it, which was
210
strongly occupiedby the military, but one whoseeconomywas little developed
andwhich was incapableof maintainingits hugearmy on its own resources,they
were interestedin employing the wealth of the economicallybetter developed
provincesfor the maintenanceof the army of Illyricum.
The clearanceof the Gothsfrom Illyricum wasprobablynot completedbefore
270. Aurelian, who in the meantimehadgoneto Italy, hadto returnsoonto repel
a dangerousattackon Pannoniaby the Vandals.I 39 Possiblythe Vandals desired
to settlein the empireafter having beenexpelledby the Gepidae.This invasion,
costing the emperora heavy battle and expenditureon food suppliesfor the
retreatingVandals,seriouslydamagedthe province of Illyricum, which during
the previous decadehad beenable to enjoy peaceand which towards the end
of the reign of Gallienushad showedsigns of a consolidation.I40 The next year
(271) whenAurelianmarchedthroughPannoniaandthe Balkanstowardsthe east
agair.stZenobiahe had to fight with the Gothswho had invadedIllyricum and
Thrace.He gaineda greatvictory,I4I but was presumablyable also to convince
himself finally that disturbancesbeyond the frontier causedby the Goths and
Gepidaecould be broughtto an end only by radical measures.He now decided
to undertakethe difficult operationof abandoningto the Goths and Gepidae
the bulwark of Dacia,which hadlong ceasedto play its part in the defenceof the
empire.This was not a sacrifice made by the Illyriciani on the altar of empire.
The abandonmentof this province, as we have seen,had been seriously con-
templatedfor a certaintime already,and when Aurelian orderedthe evacuation
presumablyfew troopsand a much diminished civilian populationremainedto
be transferredto the right bank of the Danube. Thus Illyricum was able to
obtain temporary relief from the great pressureof the barbarians;the em-
pire as a whole gainedonly an indirect advantage.It was indeedtime to solve
the problem of the Danube frontier. The provinces of Illyricum-Pannonia,
(particularly in 258-60 and in 270), Moesia Superior (in 268-70), Moesia In-
ferior and Thrace(strictly speakingsince the forties of the third century until
270)-hadhad to suffer so much that there was a danger of losing Illyricum
with its irreplaceableman-poweronce and for all.
The abandonmentof Dacia divertedthe pressurewhich had hitherto weighed
on the DanubianandBalkanprovinces.The GothsandGepidaebeganto quarrel
about the evacuatedprovince; the Goths and someof their allies emergedvic-
torious.I4Z The Gepidaehad to give way to the pressureof the Gothsby moving
westwardsand south-westwards, which createda tensionin Sarmatianterritory.
The consequences of this new situation, however,showedup only ratherlater,
when the wars againstthe Sarmatiansbeganalong the limes of Pannonia.
The other result of the evacuationof Dacia was the establishmentof a new
2II
Crises on the Danube: the rise oj the Il(yrican soldiery
provinceon the right bank of the Danube.I43 This provincewas formed by cuts
from the territory of Moesia: the north-easternpart of Upper Moesia and the
western part of Lower Moesia were detachedto createit. Legiones V Mace-
donica and XIII Geminawhich had previously beenmoved out of Dacia were
given fortressesthere, and the Romanizedpopulationof Dacia was evacuated
to settle in it. The fact that the lower Danubelimes was given two new legions
showsthat dangerfrom the Goths was reckonedwith for the future. The new
legionaryfortresses(Ratiariaand Oescus)were chosenso as to bar the two most
importantapproachesto the Balkan peninsula:the roadleadingfrom Ratiariato
Naissusand on into part of the Balkan peninsula,and the way leading from
Oescusthrough the Isker valley to Thraceand the Aegean.
212
Chapter7
The second age of prosperity: rise and collapse
The towns oj lTpper Moesia

A fresh period of active provincial policy had already begun under Marcus.
As was shownin ti.:e previouschapter,the rich mines on Mount Kosmaj south
of Singidunumwere openedunder him, probably even before the outbreakof
the Marcomannicwars, and were garrisonedby troops from Upper Pannonia.
Municipal status was given by Marcus to all the places of importancein the
valley of the Morava in Lpper Moesia along the road from Byzantium to
Singidunum(Fig. 37, p. 220). This road was later to becomeone of the most
importanttrunk roads of the empire. The exact date of thesetown foundations
is unknown, and so we cannotdecidewhetherthesemeasuresof Marcus were
inducedby experienceresulting from the war, or whetherhe had certainplans,
independentof the crisis, which he then was able in part to realize, despitethe
difficulties createdby the war. In Pannoniathe measurestaken seemwithout
exceptionto have their origin in the experiencesof the war, but it is uncertain
whetherthey were initiated by Marcus or only later by his son.! The openingof
the minesat Mount Kosmaj fits well with the systematicdevelopmentof natural
resourcesin Upper Moesia, which had beengradually carriedforward sincethe
time of Trajan.
Marcus' programmeof municipalizationhas a numberof details in common
with the Flavianfoundationof municipiain the valley of the Save(Neviodunum,
Andautonia, Siscia, Sirmium). The tendencyto createbasesand route-centres
for traffic, on a road the importanceof which was recognized,is unmistakable.
21 3
The secondage oj prosperity: rise and collapse
The towns of Naissus and Margum were certainly creationsof Marcus; the
foundation of the third town in the valley of the Morava, Horreum Margi,
cannotat presentbe dated accurately,but its foundation by Marcus would be
the most likely possibility.
Naissus was situated not far from the confluence of the Morava and the
Nisava,2in the areawhereall importantroadsthroughthe Balkansof necessity
meet: first, the Singidunum-Byzantiumroad, the principal route in the Balkan
peninsula,second,the road from the Adriatic coastat Lissus via Ulpianum to
the valley of the Morava, and, third, the road connectingNaissusvia the Timok
valley with the Danube; of this the terminal points were Ratiaria on one
bank, and Dacia on the other. The Lissus-Naissusroad was joined by the one
from Thessalonicain the Vardar valley via Scupi to the Danube.The successur
of Naissus,modern Nis, is also one of the most important route-centre'>for
Balkan traffic. Under Septimius Severus,Naissus is named as the birthplace
(origo) of a praetorianin Rome, and it must thereforealreadyhave beena town
with municipalstatus.3 From Naissusvery few decurionesareknown,b'Jt from the
fact that they were Aurelii 4 it can be concludedthat the council was composed
of native-bornnew citizens. Nor is there a large number of recordedforeign
settlersin and aroundthe town. It lay on the northernbank of the Nisavain the
region of the Turkish fortress (Kale) of Nis; one quarter,however,lay on the
southernbankandwas connectedwith the centreof the town by a bridgeacross
the Nisava. The bridge was situatedin the sameplace as the modern bridge,
nearthe southerngate of the Kale. Otherwisewe know very little of the town.
Under the high mound of the Turkish fortress the Roman layers are buried
severalmetresdeep,and the excavationsso far have done no more than estab-
lish the fact of Roman settlement.The territorium of Naissuswas remarkably
large, probably the largest of the territoria in Upper Moesia. It included not
only the valleys of the Nisavaand of the two armsof the UpperTimok (Svrljiski
and Trgoviski Timok) but also the middle valley of the Morava betweenthe
Klisura and Horreum Margi. To the west the valleys of the Jablanicaand the
Toplica belongedto Naissus,tOO.5 No urban centressprang up in all these
regions, with the exceptionof Remesiana(Bela Palanka)in the Nisava valley,
which was of specialimportanceas the seatof the Conciliumprovinciae, the place
where the Ara Augusti Provinciae was situated6 -and later as a bishop's see.
Hitherto, however,we haveno evidencefor the municipal statusof Remesiana.
Discussion of the problems connectedwith the birthplace of the Emperor
Justinian is beyond the scope of this work. This place in the valley of the
Jablanicanear Leskovacwas chosenfor the seat of the vicar of Illyricum and
was developedinto a town under the name JustinianaPrima. Recent Jugo-
214
slavianwork has solvedthe controversyabout Justinian'sbirthplaceby excava-
tions on the CariCin Grad near Leskovac.7 For problems of earlier times,
however,informationfrom the time of Justinianis of first importance,sincethis
providesthe only meansof determiningthe territorial division of UpperMoesia
in the third and fourth century. The size of the territory of Naissus,also, can
only be establishedby meansof early Byzantine sources,and below we shall
haveto refer to thesesourcesfrequently. At the latestunder SeptimiusSeverus,
a road-station(praetorium) was establishedat Naissusfor imperial officials travel-
ling on official business;this station was administeredby the stratores of the
governor of Upper Moesia. These stratores have left an imposing number of
altars.8
A strator is also attestedin HorreumMargi.9 This town is situatedat the point
wherethe road down the valley of the Morava crossesthe river. From Naissus
down-streamthe road lay on the right bank of the Morava, but near Horreum
Margi (Cuprija) it changedover to the left bank,and then, graduallyleaving the
valley, it crossedthe hilly region of north-west Serbia to Singidunum. The
crossingof the Morava was by meansof a bridge, the piers of which are still
partly preservedin the bed of the river. HorreumMargi, too, was a road junc-
tion: the greatdiagonalroad acrossthe Balkanshadto makea detourin orderto
reachthe seatof the provincial administrationof Upper Moesia, the legionary
fortress of Viminacium. The Horreum Margi-Viminacium road also left the
swampyvalley of the lower Moravasoonafter HorreumMargi, crossingover to
the valley of the Mlava.IO As a nodal point, Horreum Margi had obviously
playeda certainrole from very early on: at leastits nameshowsclearly enough
that the place had beena supply-basefor the Danubianarmy at an early date.
Of the town itself hardly anything is known. A rectanglemeasuringc. 300 by
350 m, whose walls are, however, late Roman or early Byzantine, could poss-
ibly be the successorof earlier walls.H The small area(II ha) enclosedby these
walls would correspondwell to the size of a small town. The personsknown
from the few inscriptionsfrom Cuprija were retired soldiers or civilians, all of
them originatingfrom the native population.For the existenceof a municipium
at Horreum Margi we have two pieces of evidence, one from the time of
Caracallaand the other dating to the year 224. The foundationthus falls before
Caracalla,either in the time of Marcus or in that of SeptimiusSeverus.I2
The valley of the lower Morava is very boggy even today. Becauseof its
swampinessthe roadswhich followed the valley left it nearHorreumMargi and
ran, as we have seen,through the hilly region to the Danube.Nevertheless,a
route-centrearoseat the mouthof the Morava on the actualbankof the Danube,
which had a certainimportanceas early as the secondcentury.Margum (Orasje
21 5
near Dubravica) was the seat of a conventuscivium Romanorumin the second
century,for its curator is mentionedon an inscription.I3 This showsthat Roman
citizens early settledat the mouth of the Morava, presumablyin responseto the
growing use of the river for the transportof goods, after the mines of Upper
Moesia had been put into operation. The raw material of the mines could be
transportedto the Danubeonly via the Morava. Recentexcavationshave dis-
closeda small part of the town on the bank of the Morava,I4 but we have no
knowledgeof its size and appearance.Whetherit had an area exceeding60 ha,
as older recordswould have it, can be neither proved nor disproved.The date
of its foundation can be deducedwith certainty from its name, municipium
Aurelium Augustum.I 5 The very small body of inscriptionsfrom the site are of
foreigners, mostly freedmen or slaves, and especially of Augustales,and this
agreeswell with the fact that the settlement,as seatof cives Romani,had already
beenof someimportanceearlier as a centreof merchantsand businesspeople.I6
Mines in Upper Moesia

The opening of the mines on Mount Kosmaj,I7 too, was the work of Marcus,
and presumablydatesfrom before the outbreakof the Marcomannicwars. We
have alreadyseenthat, for the protectionof this mining territory, troops were
transferredfrom Upper Pannonia. The inscription of a tribune of cohors I
Ulpia PannoniorumI8 dates from the year 168 at the latest; somewhatlater,
thesetroops were replacedby the cohortesAureliae,newly recruitedby Marcus.
The mines on Mount Kosmaj show an exceptionalfeature amongthe mines of
Upper Moesia, which suggeststhat a new policy was employed towards the
mines of the province which may have beenthe work of Marcus. This feature
is the fact that the inhabitantsof this mining territory were to all appearance
wealthierthan thoseof others,as is strikingly demonstratedby the monumental
remainsof the Kosmaj area. From the other mines of Upper Moesia either no
inscriptions at all (Aelianum, Aurelianum) or very few only (Ulpianum, Dar-
danicum) are known, whereasthe Kosmaj area has yielded a great number of
inscriptions which differ in somerespectsfrom the body of inscriptions of the
province. Thesedifferences,also, allow somedeductions.I9
First, the whole body of inscriptions from the Kosmaj area belongs to the
period from Marcus roughly to SeverusAlexander,though the majority prob-
ably belong to the early years of the Severi. In this period very good stone-
masonsworked there, who had come either from Viminacium or from some
town of the Save valley and who producedrichly ornamented,probably very
expensivegrave-reliefs of the south Pannonian-Noricantype. This at once
216
allows us to infer a suddenlydevelopedprosperity which enticedgood stone-
masonsto come there. Second,practically without exception,foreigners only
are namedon the inscriptions, e.g. Dalmatians,20Thracians,21south Moesians
and somefrom the westernprovinces,22too, a colourful medley which other-
wise can be proved to have existed on the Danube in the Flavian-Trajanic
period only. There is a remarkablenumber of peregrini among them, con-
sidering that in Upper Moesia in particularperegrinivery rarely set up inscrip-
tions.';Third,the inscriptionsare amazinglyfull of mistakes.Thereexists hardly
any text that is not incorrect in one way or another,and several are entirely
incomprehensible.All this can only mean that, under Marcus or a little later,
excellentopportunitiesfor profit were availableat the minesof Mount Kosmaj
which attracteda colourful mixture of foreigners. The prosperousconditions
attractedhither good stonemasons,too; but the patrons-partof them pere-
grini--were either uneducatedpeople or people who did not understandLatin.
The erection of inscriptions had become a fashion in which some even in-
dulged who were incapableof readingand understandingtheir own texts.
The reasonsfor this economicblossomingarenot known. One might suggest
that the outputof the newly openedmineswas abundant,or that Marcusgranted
the miners particularly favourable conditions. Neither reasonseemsunlikely,
sincea new mine with rich resourcesof lead and silver situatednot far from the
Danubeand from the road-centreat Sirmium was apparentlyconsideredby the
governmentto be an economicassetof first importance;and in that caseit will
not be surprisingto find theseparticular mines protectedby auxiliary troops.
The economicprosperityof the Kosmaj areawe can easily trace back to local
factors. A similar prosperity,observablein the Danubianprovincesin general,
hascausesthat arefar more complexand to be interpretedonly when studiedon
an imperial scale. These causeshave already been discussedin the previous
chapter. Reconstructionand the awakening self-confidenceof the Danubian
military, who playedthe leadingpart in the strugglefor power under Septimius
Severus,led to an economicupsurgein the Danubianprovinces,and especially
in their border area; this dependednot merely on the prosperity of the local
economybut also on resourcesdrawn fromthe wealth of the empireas a whole.
We are as yet far from being able to define at all accuratelythe mechanismof
this process,but must now attemptto characterizeits main features.
Urban developmentin the third century

The first certain indicationsof the intention of the Severi to follow an excep-
tional policy towards Pannoniacan be seensoon after the accessionof the first
21 7
The secondage ~f prosperity: rise and collapse
Severanemperor. Carnuntumand Aquincum, the two seatsof provincial ad-
ministrationin Pannonia,were grantedthe rank of coloniae in the year 194,23a
distinction previously bestowedonly very rarely; Hadrian, for example,found
it difficult to understandwhy towns should apply for this rank. Whether the
title of colonia in the caseof Carnuntumand Aquincum involved also a grant of
the ius Italicum or something similar, we do not know. One very important
change did, however, take place with the granting of colonial rank, and this
changeshouldbe interpretedas a favour to the settlementsoutsidethe legionary
fortresses.As far as canbe seen,it did not affect CarnuntumandAquincumalone
but all settlementsoutside the fortressesof the Danubein so far as municipal
self-governmentexistedamongthem already.
As we have seenin Chapter5, at thoselegionary fortresseswhich were also
the seatsof governorsthere had existedsince at least the time of Hadrian two
autonomouscommunitiesside by side: the settlementaroundthe fortressitself
(canabae)and distinct from it, and situateda little distanceoff, the mur,icipium
Aelium (Fig. 22). Both had local autonomywith the necessarycorporationand
magistrates.Under Septimius Severusthe canabae ceasedto exist as separate
autonomousunits andwere either given a municipaltype of constitutionof their
own, as at Dacian Apulum, or they were incorporatedinto the municipium.
Henceforthonly one municipal communityexistedat theselegionaryfortresses
andthis incorporatedthe settlementaroundthe fortress.Only at Aquincum and
Carnuntum,the seatsof the two Pannoniangovernors,is it attestedthat the
community thus enlargedwas at the sametime given the rank of colonia; the
other municipia still retainedtheir old rank for semetime, like Viminacium,
which became a colonia only under Gordian II1.24 It is remarkable that at
Viminacium, accordingto building-inscriptionsof SeptimiusSeverusand Cara-
calla, the canabae continued to exist; these emperors ordered the canabae of
Viminacium to be rebuilt.25 The stoneslab on which this is recordedis, however,
remarkablysmall, and from this we may infer that the newly built canabaewere
not comparablewith the very extensivesecond-centurysettlementad legionem.
This conclusionis supportedby the fact that we haveno epigraphicproofs at all
of the existenceof decurionesor magistratesof the canabaefor the time of the
Severi. The canabae,therefore,ceasedto exist as a type of community, and only
a very small settlement,possibly consisting of only a few houses,continued
under the control of the legion.26
It is not possibleto say why this changewas necessary.It may hold true in
generalthat SeptimiusSeverusdiminished,or evenpractically abolished,those
lands in the frontier provinceswhich were under military control, the reason
being that these territoria were no longer able to securethe provision of the
218
The secondage ofprosperity: rise and collapse
army. This emperor,being an innovator in the field of army provision, intro-
ducedthe annonamilitaris as a specialtax mainly on natural produce,which was
very oppressiveon the civilian; this must imply radical reform of the old system
of army supply.27But thesereformsdid not necessarilyinvolve the municipaliza-
tion of the canabae,sinceevenunderHadrianthey had only a partial andindirect
place in the provision of the army. Presumably,other reasonswere decisive;
we may suggestthe permissionto marry, which was the legal recognitionof a
situation which had long existedde facto. As long as the canabaeremainedas a
special type of community, the membersof soldiers' families were in theory
citizens of anothercommunity, and possibly one where not even their grand-
parentshad beenliving. The municipalizationof the canabaeabolishedthis com-
plex legal status of the soldiers' families which did not correspondto the
actual situation. Thereafterthe legal status correspondedto the true one: the
soldiers' families were citizens of the community in which they had settledin
former days. At the sametime the municipalizationmeantan improvementof
position, sincethe canabae,despitetheir pseudo-self-governing type of constitu-
tion, were not properly autonomous,in that they could not be the origo of
Romancitizens.28 It is possible,too, that betweenthe two communitieswhich
existed after Hadrian in the vicinity of the legionary fortressestensionsarose
from various reasons.As communitiesunder military control, the canabaewere
deprivedof the glamour of independence, but in somerespectsthey were privi-
leged,andamongtheir inhabitantstherewerea few rich peoplewho shouldhave
offered their riches in the neighbouringmunicipium for purposesof that com-
munity.
Before enteringfurther into the foundationsof towns by the Severi, mention
mustbe madeof somemeasurestakenby SeptimiusSeverusin Pannoniantowns,
which suggestfavour towards Pannonia.The colonia Siscia was henceforth
called colonia Septimia Siscia;29 this suggestssome significant measurethere.
No details are known, but we might supposesomereform of the financesof the
town, the grant of certainprivileges or somesimilar act. Somethinganalogous
was undertakenin Savaria, where a Kalendarium Septimianumis attested,ad-
ministeredby a vilicus, who was a municipal slave.30 The word Kalendariumper-
haps denotesa particular kind of book-keeping,presumablyof special funds
donatedto the town of Savariaby SeptimiusSeverus.
Further foundations of towns by Septimius Severusare not unequivocally
attested;but this by no meansindicatesthat he did not createany new municipal
communitiesin his favourite country. After the secondhalf of the secondcen-
tury, however,mentionof the tribus of a citizen on inscriptionsbecomesexceed-
ingly rare, and the imperial by-nameof the town, too, is mostly omitted from
21 9
Julio - Claudian Flaviansand

period Trajan
Marcus andthe
Hadrian
3 rd century
km km km
5.?O km
Figure 37 The progressof urbanization
220
The secondage of prosperity: rise and collapse
inscriptions.The foundationdate of a town can be learnt from its being named
on a dated inscription. Since these dates, too, merely provide a terminusante
quem,it is best to enumeratehere all foundationsof the third century, because
everythingsuggeststhat the provincesof Pannoniaand Moesia Superiorwere
entirely municipalizedby the end of the third century-thatis to say,their entire
area consistedof municipal territoria (Fig. 37). Since the majority of the new
towns are namedas municipalities for the first time under Caracallaor a little
later, it is very probablethat they were promotedto the rank of municipium
underCaracalla,presumablyas a result of the Constitutio Antoniniana.As a result
of this enactmentthe civitatesperegrinaeas such vanishedbut continuedto exist
as de facto municipal communities;the momentat which they becamea munici-
pium in the formal sensepresumablydependedonly on the ability to assemblea
town council (ordo).31 At any rate we do not know of any non-municipalcom-
munities after Caracalla,32and everything suggeststhat sooner or later-but
most of them during the time of the Severi-all civitates were convertedinto
real municipia with decurionesand magistrates.The incorporationof the canabae
in the adjacentmunicipium or colonia, whicheverit might be, alreadyillustrates
a tendencyto abolishdifferencesbetweentypes of communities,in other words
to createa common denominatorfor the various territorial authorities. This
trend led to the very simplified system of the late Roman period, when all
autonomouscommunitieswere called civitas, and words like municipium and
colonia becameso devoid of meaningthat they were hardly usedany more, for
instancein the Codex Theodosianus.33
In PannoniaSuperior, Vindobona andBrigetio, the two other places with

legionaryfortresses,probablydid not sharein the improvementof statusunder
Septimius Severus.Brigetio is named MUllicipium Brigetio Antoninianumon an
inscription,34 and a foundation by Caracallacan be inferred. Certainly by 217
Brigetio was a municipium.35 At that time, or possiblylater still, Vindobonaalso
becamea municipium; of its decuriones and magistrates,only one is as yet
known.36 In Vindobonaas well as in Brigetio the duality of settlement,described
in Chapter5, was probably completelydevelopedbefore the end of the second
century, but the settlementlying further away from the fortress was given
municipal statusonly in the third century; thegrant, however, will have also
included the canabae in this improvement of rank. The pseudo-autonomous
organizationsof the canabaeat any rate ceasedto exist in Brigetio, too, and some
of the decurionesof the municipium Brigetio had themselvesburied in the former
canabae,and furthermorethey erectedbuildings there; from this it must follow
that they were citizens and inhabitantsof the settlementaround the legionary
fortress.37
221
Further municipal communities of Upper Pannonia which obtained this
statusprobablyonly during the third century were Jovia on the Drave east of
Poetovioand the municipium Faustinianumsomewhereeastof Siscia.If these
towns are enteredon a map togetherwith the rest of the earlier foundations
(Fig. 37) it can be seenthat only one areaof this provincedoesnot containany
municipium known to us: this is the areasouthof Lake Pelsowherewe located
the Cotini in Chapter3. Here only Tricciana(Sagvar)is a possiblecandidate,but
up to the presentno proofs of its municipalstatusareavailable.Furthermoreit is
not certain that Tricciana lay in Upper Pannonia.
Jovia, probably known originally as Botivo, is mentionedin the Itinerarium
Burdigalense(Hieroso!Jmitanum)as a civitas,38showingthat it must havebecomea
town at the latestunder the Tetrarchy.No more is known aboutit, exceptthat
it was a bishop'sseein the fourth century.39
The municipium Faustinianumis mentionedas the homeof a praetorianand
an equessingularis in Rome;40 in addition, its scriba is namedon an inscription in
Siscia.41 The find-spot of this inscriptionitself is sufficient to suggestthat this
municipium must be soughtsomewherein the vicinity of Siscia. And sincethe
valley of the Saveabove Siscia was alreadycompletelymunicipalized,only the
valley below Siscia comes into question as its territorium. Siscia incidentally
furnishes a further illustration that in the small Pannoniantowns the scriba
was of high social position and as the de facto director of municipal affairs was
presumablyfairly rich, too. The scriba of the municipium Faustinianumwas at
the sametime Augustalisin Siscia.
In PannoniaInferior, the only new town to be mentionedis Mursella on the
Drave west of Mursa. This town is referred to on a single inscription only,42
which mentions a decurio M. Aurelius. However, its foundation is not easily
understandable:it was only ten or twelve miles from the Hadrianic colonia
Mursa, and-possiblyfor this very reason-itis not recognizedas a town in the
Itinerarium Burdigalense, which specifies the ranks of places (civitas, mansio,
mutatio). Moreover,thereis a possibility that the decurio on the inscriptionwas in
fact a town-councillor of the town of the samenamein Upper Pannonia.If so,
Mursella (Mursa minor) in Lower Pannoniashould not be countedamongthe
municipia.
The map of Lower Pannoniashows more gaps than that of the Upper pro-
vince. In the areaof the civitas Breucorumno urbancentreis known as yet, nor
is a municipium known in the areawhich was formerly the civitas of the Her-
cuniates.The municipium Volg (...) could be claimedfor the Hercuniates,but
it is entirely unidentifiable,since it is only namedon an inscription from Inter-
cisa.43 As a town with municipal status,Sopianae(Pees)on the southernslopeof
222
the Mecsekmountainscomesinto consideration.As early as the time when the
Italians started to extend their trade-connectionsin Pannonia,it had been of
importanceas a route-centre,44and under Diocletian it becamethe seatof the
civil administrationof the province of Valeria.
In Moesia Superior the civil settlementof the legionary fortress of Singi-
dunum was also madea municipium, presumablyunder Caracalla.4s The topo-
graphy of the Romansettlementsunder modernBeogradis not yet altogether
clear.46It is known that the legionaryfortresslay on the site of the later fortress
at the confluenceof Danubeand Save(Kalemegdan,Tvrdjava). Tracesof a civil
settlementon the southernand easternoutskirts of the fortress have already
beenestablished,while uncertaintracesfurther eaststill may perhapsindicate a
secondcivil settlement.Here, too, the municipium presumablycomprisedboth
settlements.
Of the remainingmunicipia of Upper Moesia,two can be datedto the period
of the Severi and presumablyto the reign of Caracalla,the only basis of the
dating, however,being the fact that their decurionescarry the nameM. Aurelius.
But as only one decurio of either of thesetowns is so far known, even greater
cautionis neededin determiningthe foundation dateof others.The two towns
wheredecurionesare known are the municipium Dardanorumor Dardanicumin
the valley of the Ibar 47 and the municipium Celegerorumin the valley of the
westernMorava. The identification of municipium Dardanorum,whose name
was long known in the abbreviationmun. D D, 48 with Socanicain the Ibar valley
becamepossibleonly recently.Here hadlain the seatof the administrationof the
metalla Dardanica. Of the municipium Celegerorumhardly anything is known.
The tombstoneof its only attesteddecurio was discoveredlong ago, but its find-
spot (Ivanjica) posesa number of problems, since the valley in which it lies
opensonly towardsDalmatia, andit would have beendifficult in this secluded
areafor a town to come into existence.The decurio may have hadhis estatesin
the valley of the Moravica, but municipium Celegerorummust be looked for
somewherein the valley of the westernMorava (perhapsat Kraljevo).
The nomenclatureof thesetwo Upper Moesiantowns recalls the Pannonian
municipia Latobicorumand!asorum.We might thereforededucethat both had
beenformed from civitatesperegrinae(of the Dardaniandthe Celegeri).But this is
true only of municipium Celegerorum,for municipium Dardanorumgoesback
to the metalla Dardanica and this is the reasonwhy we cannotexcludethe poss-
ibility that its real namewas municipium Dardanicum.
The remainingthreeUpper Moesiantowns were also basedon former mining
districts. SouthofViminacium in the valley of the Mlava the Itineraries indicatea
station named Municipium;in this area had lain the territorium of the metalla
223
Aeliana Pincensia.4 9 In the region of the metalla Aureliana, however, we have
knowledgeof a placecalled Aureliana,50which was presumablythe centreof an
otherwiseunattestedmunicipiumAurelianum.Theassumptionthatamunicipium
existed in the area of the metalla Aureliana is justified by the fact that in early
Byzantine times a separatexcfJpa (chora) with its centre Aquae existed here,s!
and that the areaof this xcfJpa consistsof parts of the provinceof Upper Moesia,
which cannot be attributed to any of the known urban territoria. 52 Since early
Byzantine territories (of the time of Justinian) are directly derived from the
urban territories of imperial times, a precursorof the chora of Aquae has to be
presupposed.By analogy with municipium Ulpianum (=metalla Ulpiana) and
municipium Dardanorum(=metalla Dardanica) we can assumethat the metalla
Aelianabecamea municipiumAelianum and the metalla Aurelianaa municipium
Aurelianum; supporting evidenceis given by the facts that a municipiu~n is
attestedin the areaof the metalla Aelianaandthat Ulpiana,the lateantiqueversion
of the nameUlpianum,correspondsin form exactlyto the place-nameAureliana.
From the order in which the place-namesof the chora of Aquae are given it can
be inferred that Aureliana was situatedsomewherenear the confluenceof the
White andthe Black Timok; this was the naturalcentreof the chora and is where
todaythe mostimportantplaceof the SerbianKrajina, Zajecar,lies. Not far from
Zajeearis a well-known Romansite, Kostol, wherea large,palace-likehousehas
already been excavated(Fig. 49) and where older notes report a fairly large
Roman settlement(Fig. 31) (of c. 16 ha).53 The municipium Aelianum was
situatedin the valley of theMlava, nearthemodernKaliSte, BozevacandBatusa.54
The third town of late foundation, also basedon the municipalizationof a
mining district, was Aureus Mons (Seona)on the bank of the Danubeeast of
Singidunum.It is listed as a civitas in the Itinerarium Burdigalense.55 Quite nearto
this otherwiseunimportanttown lay the mines of Mount Kosmaj, which most
probably were incorporatedin the territory of Aureus Mons. We do not know
whenthis occurred,but it was certainlyonly after the time of the Severi,sinceno
traceof a municipal administrationis extant on the numerousinscriptionsfrom
the period when this district of mines enjoyedits greatestprosperity.
As we have seen,the municipalizationof Upper Moesia was basedto a con-
siderableextenton the mines. At first glanceit doesnot seemprobablethat the
mining districts (territoria metalIi) becamemunicipalized, as this involved re-
moval from direct imperial administration;but a law of the late Romanperiod
refersto the UpperMoesianminesin particularas beingadministered'according
to ancient custom'56 by the curiales of the towns. On an inscription the metalla
Dardanica are called Metalla Mttnicipii, proving that the municipalizationof these
mines had already occurredin the third century.57Whether this reform of ad-
224
ministration has to be attributed to a decline in the production of the mines
cannotbe decided,but it is noteworthythat the undeniableprosperityon Mount
Kosmaj lastedonly to the municipalization.58
It is not improbablethat the settlementswhich had sprungup nearthe mines
were the only urbansettlementsin the areasin question,or more accurately,the
only centreson which a municipal autonomy could be built. This is perhaps
the easiestway to explain the municipalization of the mines. Moreover, the
town foundationsof the third century were not able to producetowns in the
true senseof the word whetherin Pannoniaor in Upper Moesia. The difficulty
encounteredin trying to determinethe numberor eventhe sites of thesetowns
is causedby the absenceof any epigraphicalor archaeologicaltracesworth men-
tioning. We haveonly to recall our ignoranceof a singledecurio at a considerable
number of these towns (Aelianum, Aurelianum, Aureus Mons, Horreum
Margi, Jovia), and the fact that the remaining towns have mostly produced
only one decurio on inscriptions (mun. Dardanorum,mun. Celegerorum,mun.
Volg ... , Mursella, municipium Faustinianum);whereasfrom towns which
truly deservethe title of 'town' a considerablenumber of decuriones can be
listed.
In this connect.ionit is a little obscurewhy the centreswhich alreadyexistedin
some districts had not obtained municipal status. To all appearance,Aureus
Mons was much less important than the mining settlementon Mount Kosmaj,
and in the areaof AurelianumTimacumminus, a civil settlementof importance
at the junction of the Naissus-Ratiariaand Naissus-Aurelianumroads had de-
velopednearthe fort of cohorsII Aurelia Dardanorum.59 In the sameway, not
oneof the more importantlimes-fortson the Danubewas chosenas the centreof
the municipalized civitas Hercuniatium; instead municipium Volg( ...) was
possibly selected,whose position though unknown was certainly not on the
Danube,since it is not listed among the places on the limes-road either in the
AntonineItinerary or in the PeutingerTable. We might possibly concludethat
the governmenttendedto grant municipal rights only to civil settlements,but
not to settlementsround forts. Aureus Mons was in fact probably situatednear
an auxiliary fort, but even more troops were stationedat Mount Kosmaj.
As well as new foundations,promotions of towns to higher rank must be
mentioned. These were apparentlybasedon entirely political considerations.
As alreadymentioned(p. 2 18) SeptimiusSeverusgrantedthe provincial capitals
of the two Pannoniasthe rank of colonia; during the time of the Severi, Cibalae
and Bassianaalso becamecolonies,60then Viminacium61 under Gordianusin
239, and Singidunumand Brigetio 62 at an unknowndatetowardsthe middle of
the third century. Whetherfurther towns were grantedthe rank of colonia is an
225
The secondage oj prosperiry: rise and collapse
open question; at any rate the promotions thus far clearly illustrate motives
basedon power-politics.First, the provincial capitals,whosetroops had helped
SeptimiusSeverusto the throne,were distinguished;not much later the towns
Cibalaeand Bassianain the Sirmian region-aproof that the significanceof this
area was clearly recognizedby the Severi; then followed the capital of Upper
Moesia, Viminacium, and finally the towns with legionary fortresses,Brigetio
and Singidunum.And this meansthat only the towns with legionaryfortresses
and towns in the surroundingsof Sirmium were distinguishedwith the rank of
colonia; all of them were towns of particular relevanceto the basis of imperial
power in the third century.
Town foundations,of course, only createdthe administrativeand juridic~.l
frame of a communallife, but the latter could never be createdsolely by the
formal procedureof foundation. Although in the third century some com-
munities of the empire-aboveall the towns proper who could demandfull
autonomy-hada preferencefor emphasizingthat they formed a res publica-
a community of highest jurisdiction-andalthough some Africans, Spaniards
or Orientals may have cherishedthe utopian view that the empire was a con-
glomerationof towns equivalent,autonomousand possessed of equal rights,63
nowhere had these ideas less foundation than in some parts of the Danube
provinces,where urban life with all its requisities could never arise. Even in
thosetownswhich as military coloniae(e.g. Ratiaria)or ascanabae(e.g. Aquincum)
producedan active communallife and becameimportant centresof trade and
industry, it can be noted that the municipal upper class and in particular the
decurionesoften owed their wealthand social positi0!1 not to participationin local
economiclife, but to the circumstancethat they representedsomebranchof the
administrationof the empire.In Ratiariathe customs-farmingfamilies of Sabinii,
Julii and Antonii played a leading part in the ordo of the colonia,64 and the
customs-farmerTitius Antonius Peculiariswas one of the richest decurionsof
Aquincum and had connectionswith the municipium of Singidunumas well. 6s
It is obviously one of the most characteristictraits of urbanlife on the Danube
that wealth and prosperity were not basedon the suddengrowth of the local
economy but rather on participation in the central power or simply in the
administration.Wealth was derived from empire-wide sourcesand not from
thoseof local production.We shall refer to this featureof the Danubeprovinces
morethanoncein this chapter;herea shortsummarywas necessary,becauseone
important characteristicof town life is particularly connectedwith this feature.
As we have seen,the majority of the newly founded towns were incapableof
producinga municipallife. The decurionsof thesetownshadneitherthe material
nor the cultural backgroundto supporteither the wish or the ability to immor-
226
talize themselveson inscriptions; only the canabaeof the legionary fortresses
figure as an exceptionto this rule. Therewas more than one reasonwhy munici-
pal statuswas first grantedto them only under the Severi, but evenbeforethis
they had beenable to producean active communallife.
Syriansin Pannonia
During the time of the Severi a very extensiveoriental immigration becomes
perceptiblein the towns in the true senseof the word which has not yet found
a final explanation. The possibility has frequently been consideredthat this
immigration was part of the programmeof reconstructionafter the Marcoman-
nic wars; organizedresettlementhas also beensuggestedas a possibility on the
suppositionthat the Orientals had not voluntarily come to the Danube. In
generalhistoriansare in agreementthat the Orientals contributedsignificantly
to the economicprosperityof the Danubeprovincesin the time of the Severi.
Intercisa,the fort of cohorsI Hemesenorum,is generallyreferredto, where,as a
result of reinforcementof the regimentfrom Syria, a Syrian enclavecameinto
existence; and this enclave, according to the abundant body of finds and
monuments,66createda prosperity unparalleledat other sites with auxiliary
forts on the Pannonianlimes. Thus arosethe attractivepictureof the reconstruc-
tion of a Pannonialaid waste in the Marcomannicwar: with Marcus already
transferringoriental troops to Pannonia,and the civilian Orientalsthen follow-
ing, inducedto migrateto Pannoniaby the governmentin orderto employtheir
long-renownedbusinessacumenand their economicability for the benefit of a
provincethat had sufferedseverely.And it is true that Orientalsare much more
often attestedin Pannoniathan in Upper Moesia, which had been much less
exposedto barbarianattacks.
To this picture somecorrectionsmust be made.First, oriental immigrationis
already detectablein the secondcentury. The decurio of Carnuntum,Domitius
Zmaragdus,who had the amphitheatrebuilt at his own expensearound the
middle of the secondcentury, came from Antioch; 67 the freed woman of a
Syrian namedBargatescan be tracedat the Danubebendas early as the turn of
the first century;68 Syriansoldiersservedon the Danubeevenbeforethe time of
Marcus,andso on. It is well known that the large-scalecolonizationby Trajanof
newly conqueredDacia brought (or ratherlured) a greatnumberof Orientalsto
that province,and this oriental immigration left tracesin the two Moesias,too,
whereOrientalsareas well attestedbeforeMarcusas afterwards.0 9 The Orientals
who reachedPannoniain the time of the Severiwere following a route trodden
by Orientalsfor at least a century, and we must not forget the part played by
227
Aquileia: this port, like her successorVenice, was the greatestgatewayto the
eastin Europe.
In the secondplace,it is presumablyno coincidencethat the Orientalsdid not
settlein the towns which needed'economicaid' most urgently: they are absent
in the small towns that had beenfoundeda short time previously or were then
just aboutto be founded,but they gatheredall the morein thosetowns that were
flourishing centresof the province anyway.
It cannot be determinedwhen the actual immigration of the Orientals into
Pannonia began at this period. The earliest immigrants are attested under
SeptimiusSeverus,excluding,of course,the soldiersandtheir families belonging
to the cohors Hemesenorum,who made their appearanceat Intercisa under
Marcus or Commodus.Thereis also the possibility that immigrationon a larger
scale startedonly when the legions of Pannoniawere employedin the Orient
againstPescenniusNiger and then againstthe Parthians.It is known that the
legions of Pannonia,and thoseof Upper Moesia too, brought oriental soldiers
with them on their returnfrom the East.70 Thesemenpresumablybroughttheir
families with them, and thus a longer period of immigration could have begun.
There was however an undoubtedimmigration of completely civilian ele-
ments,aboveall Jews,inhabitantsof Asia Minor andespeciallySyrians.Sporadic
epigraphicevidenceprovesthe existenceof larger or smalleroriental settlements
in Sirmium, Mursa and Savaria.This evidenceis for the most part confined to
tombstones,7Ibut in Mursa a Jewishhouseof prayer is attested,which was re-
constructedin the reign of Septimius Severus,72and a synagoguemust be as-
sumedto haveexistedin Intercisa,wherethe headof the synagogaIudaeorumwas a
customs official named Cosmius.73 Remarkableis the sporadic occurrenceof
Jews in north-easternPannoniain the area round Aquincum,74though inside
the town they are less well attested.A small Syrianenclavewas formed nearthe
auxiliary fort Ulcisia Castra,where in the period of the Severi a Syrian cohort
was stationed.75
As one can deducefrom the epigraphicevidence,theseenclaveslived lives
which were rather isolatedfrom the local population. The Syriansin Intercisa
married only among themselves,their sole outside links being with Syrian
families in Aquincum; the reasonfor this is mainly that somesoldiers'sonsfrom
the circle of the cohors Hemesenorumwere enlistedin legio II Adiutrix. The
Syriansin Ulcisia Castra,too, had connectionsonly with the Syriansof Aquin-
cum. This isolation is all the more striking since the Syriansin Aquincum and
especially in Brigetio had worked their way up to membershipof the ordo
decurionum,76and at the latter town became,to all appearance,the richest and
most important decurionsof the town. Aquincum and Brigetio were the two
228
largestcentresof oriental immigration, where newly arrived Orientals can also
be provedin the later yearsof the Severi.77 Nor did immigration ceaseafter the
Severanage; even in the fourth century fresh groups can be proved to have
arrived.78
The markedly rural origin of the Syrian immigrantsis noteworthy. On their
inscriptions they often state their origin with full details, not only of the city
territorium but also of the native village within it.79 The possibility might
thereforebe entertainedthat theseSyrian inhabitantsof villages had beensol-
diers who after their dischargehad madetheir way througha municipal career,
and as suchno longer recordedtheir military service.This possibility, however,
can be confidently dismissed,since the Syrian decurions of Brigetio did, in
fact, sometimesstatethat they were veterans,80and the circumstantialstatement
of the native place of origin would not be imaginablewithout an equally cir-
cumstantialstatementof the military career.Thus it seemsmore probablethat
someSyrianscameto Pannoniaas civilians, but it remainsto be explainedwhy
they came in such great numbers,and why during the time of the Severi in
particular.
As we have statedabove, Orientals, and among them Syrians, were already
representedin the municipal upper class of Pannonia and other Danubian
provincesduring the secondcentury. Relationswith the Danubeareawere thus
alreadydeveloped,and Syrian merchantswere also to be found more and more
frequently in the provincesof westernEurope.An emigrationfrom Syria was
therefore going on, but it certainly cannot be attributed to some decision of
imperial policy. Furthermore,we must note that Orientals appearedin great
numbersunder the Severi in Pannoniain particular, but not in other Danube
provinces. In these a suddenincreaseof immigration during the time of the
Severi can be proved only inasmuchas this period has yielded most of the in-
scriptions,andthereforeepigraphicevidencefor Orientalsis virtually confined
to it. Only in Pannoniais a preponderanceof Orientals in the ordo of certain
towns or more generally in the population of some of the settlementsestab-
lished for Severantimes. Thus we may concludethat the Syrians, who knew
very well how to win the marketsof the empire and to exploit economicpossi-
bilities offeredby someof the provinces,found in Pannoniaaparticularlyfavour-
able field for activity and were free to develop it. Since we are here concerned
with Pannoniain the period of the Severi we must associateoriental immigra-
tion with the economicupsurgewhich occurredthere at this time, and which
lured businessmen, andaboveall of coursethe Syrianswho werethe most active
of them all, to Pannonia.
Thus we must interpret the oriental (Syrian) immigration into Pannoniain
229
the contextof a suddengeneraldevelopment,but we cannotrecognizein them
the cause or motive power of this boom. Other immigrants are, of course,
attestedtoo, especially Thracians, Africans and, sporadically, elementsfrom
otherpartsof the empire.Thesewere,however,lesssignificant both numerically
and socially, and indeedtheir arrival has a close connectionwith certaintroop-
movements.Africans were brought to Pannoniaas early as underMarcus, and
at the sametime someItalians enteredthe legions.81 The Thracianswere for the
most part incorporatedin Pannonianunits when the latter marchedthrough
Thraceon their way to the east.82. The Thracianelementin the military was very
strong in the time of the Severi.83
Prosperi~y
in the earry third century
Immigration in the Severanperiod is only one of many symptoms of the
economicadvancewhich createdthe secondand greatestperiod of prosperity
in the Danubeprovinces,and in the province of Pannoniain particular. The
causesof this prosperity, in so far as they lay in central power-politics, have
alreadybeendiscussedin the last chapter.It now remainsto set out the further
evidence.
The most conspicuousphenomenonis the exceedingly large number of
inscriptionsfrom the time of the Severi.During the Severanperiod the fashion
of erectinginscriptionswas widespreadthroughoutthe entire empire.We must
not therefore attach too much significanceto the mere preponderanceof in-
scriptions at this time. The fact that the large number of inscriptions was,
nevertheless,a symptomof economicprosperityemergesonly from their varying
frequencyfrom territory to territory. Sinceinscriptionshave beencollectedor
at leastcopiedfor centuries,whereasthe systematicarchaeologicalinvestigation
of the ground in some parts of the Danubianarea is a comparativelyrecent
development,the epigraphicsourceshave a higher value as evidencefor the
identification of regional differences. And the composition of the epigraphic
materialis also significant: in contrastwith pre-Severantimes, votive, building-
and honorific inscriptions are very frequent, permitting deductionsof several
importantchangesin public life, in religion andin the social structure.
On the whole it is true to say that the majority of Romaninscriptionsin the
Danubeprovincesbelongto the Severanperiod. But this verdict is by no means
valid for towns like Poetovio, Savaria, Siscia, Sirmium, Ulpianum and Scupi
whose inscriptions for the most part date from earlier times.84 In Brigetio,
Aquincum, Singidunum,Viminacium, Timacum minus, Ratiaria, Intercisaand
in someparts of the border zone of the Danubianlimes, however,inscriptions
230
were first erectedin large quantities in Severantimes; while in several small
towns in the interior of the province of Moesia Superiorthe setting up of in-
scriptions was begun for the first time only under the Severi, for instanceat
Horreum Margi, Naissus, Municipium Celegerorumor in the Kosmaj area.
Inscriptions were set up regularly only after a community life in the Roman
mannerhad sprungup, and presentand future appreciationof them could be
relied upon. The formation of autonomousRoman municipalities in the in-
terior of Upper Moesia thus provides sufficient explanationfor the appearance
of the fashion for inscriptions. But in settlementswhere community life had
already been consolidatedduring the first or secondcenturiesthe suddenin-
creaseof inscriptionswill haveto be explainedin different terms.
To start with it is no mere chance that the accumulationof inscriptions
particularly attestsprosperity in those settlementswhich, becauseof the pre-
sence of a fort, had military inhabitants. In the towns in the interior of the
provinces,containingan entirely civil population,this economicboom cannot
be traced,at leastnot with the aid of inscriptions;and this is anotherindication
of the characterof the boom. Here we cometo the secondquestion,which faces
us when we considerthe compositionof the body of inscriptions. As already
mentioned,the epigraphicmaterialof the Severanperiod is remarkablyvaried:
it consistsnot only of inscriptionson tombs-thoughthesewere in the majority
in the periodbeforethe Severi-butof a considerablenumberof altars,building-
inscriptions and various other kinds of inscriptions as well, all of which were
rare in the pre-Severanperiod. In earlier times altars and other votive inscrip-
tions were,in fact, erectedonly exceptionally:thosesettingthemup were mostly
town magistrates,governorsandthe personnelof theportorium. This fact allows
the conclusionthat thesevota were official monuments.The deities honouredon
these inscriptions were gods and goddessesof the official Roman pantheon;
local traits did not enterthe dedicationsor were official too, suchas Danuvius,
Diana and other gods honouredby imperial officials. In sometowns this situa-
tion remainedunalteredin the Severanperiod, for instanceat Scupi.85 Else-
where, however, particularly in the towns along the Danubelimes, altars for
gods not previously attestedin the province becomefrequent; the pantheon
that can be tracedin the inscriptionsbecomesvery colourful and displayslocal
traits too, while among those setting up altars practically all social layers are
represented:there are huge altars and imposing basesof statuesand very small
primitive altars,presumablyproducedby the personsettingthe altar up himself,
with scratchedwriting of cursive character.The building inscriptions point in
the samedirection. Most of theseare dedicationsfrom temples and shrines86
built by a very varied selectionof the population. Their concentrationin the
231
time of the Severimustbe partly attributedto a contemporaryfashionfor setting
up inscriptions. A respectablenumber indicate not new buildings but recon-
structions of buildings, most of which are presumablypre-Severanin origin.
These buildings had originally lacked an inscription; building-inscriptionsof
pre-Severandate are very rare. Nevertheless,to some extent it is possible to
deducea lively building activity from the large numberof Severaninscriptions,
and this is corroboratedby direct archaeologicalevidence.
The inscriptionsof Severantimes thus attestan undeniableprosperityamong
the provincial populationof the frontier regions.A large part of the population,
including simple soldiers without rank or office, were able in this period to
afford the erection of grave-inscriptions(Fig. 38), and they were capable,too,
of displaying building activity and a certain munificence;but the fact that this
prosperitydid not extendover the whole territory of the provincefollows from
there being hardly any tracesof it at all amonginscriptionsfrom the inner parts
of Pannoniaand Moesia. Moreover, the suddenlyincreasedquantity of inscrip-
tions shows traits that allow deductionsabout important changesin society.
To startwith we haveto draw attentiononcemore to the votive andbuilding-
inscriptions. As already stated, dedicationsand buildings of private character
becamefrequent. A consistent distinction, however, between private and
official dedicationsis neitherpossiblenor desirable,for thereexistsan imposing
number of votive inscriptions, set up for the welfare of the emperoror some
regimentor corporation,the erectorsof which were private personsor ordinary
soldiers. All votive dedicationshave somethingpublic and impersonalabout
them, and so, if votive and building-inscriptionsbecomefrequentin general,an
intensive participation of wide classesof society in public life can be deduced.
From this point of view, too, it will not be possibleto considerthat their con-
centration in the frontier area is entirely accidental: the politically favoured
soldiery of the Severi were stationedin this zone which, as we have shown,
was also possessed of material privileges. The fact that this rapid development
can be demonstratedmuch betterin Pannoniathan in Upper Moesiaprovesthe
economiccharacterof its causes.Among the Danubianarmies the Pannonian
military forces were the most privileged since from the view-point of power-
politics they possessed the greatestimportance.In the undeniablylarge body of
inscriptions from Singidunum or Viminacium, ordinary soldiers or simple
private personsare much less representedthan in Pannonia.87
We must now consideraspectsof religious history which are revealedby the
dedicationsof the Severanperiod. To the economicaland political prosperity
of the frontier areais due the appearanceof local traits in religion. The earliest
indicationsof thesedateback to the first year of SeptimiusSeverus'reign, when
23 2
1st - 2nd century
srd
srd
srd
• Tombstones
+ Place- names
-IANACE)
srdcen'bvry
o 100 200
km
Figure}S Distribution of tombstonesin Pannonia
233
the Matres Pannoniorumwere honoured at Lugdunum, and the Pannonian
augurs announcedthe victory of the new emperor.88 The increasingpolitical
activity of the Pannonianmilitary thus also finds its expressionin the numerous
votive inscriptions. The strictly Roman form is always used, and this is con-
nectedwith further characteristicsof the inscriptions of the Severanperiod.
As we have alreadyseenin Chapter5, grave-inscriptionsare known in great
numbersfrom someparts of Pannoniaand from the Metohija. They were erec-
ted by native inhabitantswho mostly carriedlocal, non-Latin (Celtic, Illyrian or
Thracian) names.In the area of distribution of these early-imperial funerary
inscriptionsgrave-inscriptionsare equally frequentin the period of the Severi;
with negligible exceptions,however, the personalnamesare colourlessLatin
ones. A few Celtic names,89which exceptionallyturn up among thefamilies
which erectedthesetombstones,showthat thesewere descended from the native
families which hadearlier beennamedon tombstonesat the very sameplace.But
now the local traits of native culture, which hadbeendeliberatelycultivatedand
preservedin the secondcentury,were abandoned;a Latinization of nameswas
taking place which went hand in hand with the abandonmentof further local
customs,suchas of wagon-graves,tumuli and suchpeculiaritiesin the sculpture
of tombstonesas astral symbols,in the samearea.Sincenomenclaturewas only
rarely and sporadicallyLatinized in the secondcentury,we mustrecognizein its
suddenLatinization an important process,which, in the last resort, was based
on the desireof the indigenouspopulationconsciouslyto adaptitself to Roman
forms.
A comparisonof the zoneswhere inscriptions of the first and secondcen-
turies have beenfound, with those in which those of later dateare distributed,
immediately shows (Fig. 38) that they are roughly identical, the difference be-
ing that in the southand west, that is in the interior of Pannonia,the find-spots
becomemore scattered,whereasin the frontier areasthey are either equally or
possibly even more frequent than they were before. But it is curious that the
stationsestablishedfor beneftciarii,90 whosenetwork was developedpreciselyin
the early Severanperiod, can be demonstratedonly in thoseareasof the country
where no or very few inscriptions were erected(Fig. 39). The security service
was presumablymost necessaryin thoseplaceswhere it was to be feared that
inner tensionswould lead to movementsof unrest and robbery-bothcalled
latrocinium.
The prior condition for a larger numberof tomb-inscriptionsis a social class
that is strong in numbersand in wealth. Where grave-inscriptionsare rare or
altogetherabsent,the mostprosperousuppersociallayer wassmall in number;in
otherwords,societywasin its economicaspectstronglypolarized.It cannotwell
2H
srd
srdsecond
srd srd
second
srd
second
second
o' 100
.r::'
200 km
Figure J9 Stationsof beneftciarii
be accidentalthat thosepartsof Pannonia,wherein the first andsecondcenturies

numbersof grave-inscriptionswere set up and nonein the third century,were,
all of them,areaswhereeitheran early colonizationor the developmentof great
estatesin the secondcentury can be demonstrated,Among thesemust be in-
cluded the surroundingsof the town of Savaria,the district to the north of
Poetovio,the territory of Scarbantia,and to a certain degreealso the neigh-
bourhoodof Neviodunum.North of Lake Pelsograve-inscriptionsdo, in fact,
235
occur in the third century too, but in contrastto the previous period, when it
was alien settlersfor the most partwho set up inscriptions,theseare erectedby
soldiersburiedin their own nativevillages.91 Thesetombstonesrecall thosefrom
the villages of the frontier area,which are mainly family gravestonesof soldiers
of an adjacentgarrison.
Accordingly we must attributethe cessationof tombstonesin inner Pannonia
to the establishmentor further developmentof great estates;and the large
number of funerary inscriptions in the frontier area and north of Lake Pelso
must be attributed to the comparativelyfavourable material position of the
soldiers.One might even hazardthe conclusionthat a certainprosperity,or at
least a stateof affairs that expressesitself in the numberof inscriptions,can be
proved only in placeswhererecruitmentswere madefrom the local population.
Admittedly, the towns of inner Pannoniacontinuedto provide recruits for the
legionsand auxiliariesin the third centurybut thesesoldiersare eithernot trace-
able on inscriptions,or they left behind inscriptionsin the towns only and not
in the country. It is thereforeperhapsnot unlikely that the soldierswho origin-
ated from inner Pannoniawere mostly inhabitantsof towns. A final answeris
not yet possible,but the privileged position of the frontier areaswhich even
in the secondcenturyhad provided the greatestnumberof recruits seemsto be
certain.
In Upper Moesia the situation shapeda little differently. Until the Severan
period the governmentrefrained for security reasonsfrom making use of the
population of the whole of the province. In the secondcentury recruitments
were madeonly in the two military coloniaeand only for the legions.Marcuswas
the first to form someauxiliary troops from Dardanians;approximatelyat the
end of the secondcentury, recruitment,for the legions as well as for auxiliaries,
startedin other parts of the provincethough possibly still not all. By meansof
soldiers' recordsof their native place, and by meansof the family gravestones
of soldiersreturnedhome or buried in their native villages, recruitment,which
probably took place regularly, is attestedin the valley of the Morava (Naissus,
Horreum Margi) and in the Dardanianarea (Ulpianum). It is however very
characteristicthat at HorreumMargi 92 and in the territoryof Naissus93 erection
of grave-inscriptionsbeganonly with the recruitments,and the inscriptionsare,
with few exceptions,all gravestonesof soldiers'families.94
To sum up: the epigraphicmaterial, the majority of which datesto the time
of the Severi, is mostly concentratedin the frontier area,proving an economic
prosperitywhich is closely connectedwith the military. The compositionof the
body of inscriptionsallows us to deduceincreasingparticipationby the popula-
tion in public life; this activity causedan adaptationto Romanforms and for-
236
malities, and the abandonmentof local, non-Romancustomsin nomenclature,
and in burials and the like.
Economic prosperity left manifold traces in the archaeologicalfinds of the
frontier areas.Although excavationshave,so far, beenmadein few of the towns
on the bank of the Danube anddespitethe fact that the stratigraphyof the sites
was not always satisfactorilydealt with, one can say in generalthat stonebuild-
ings and certainelementaryrequisitesof cultivatedRomanlife obtainedgeneral
currency in towns no earlier than Severantimes. The most striking ground
plans in the parts of Aquincum laid openso far and presumablyof Carnuntum,
too, are Severanin date, and this applies in particular to the private buildings
and small shrines (Figs 28, 30). In the canabae of Viminacium the best-built
housesseemto originatefrom the sameperiod.95 The decurionsof the Danubian
towns were remarkablygenerousin the first third of the third century; a large
numberof inscriptionsattestvarious buildings, donationsand endowmentsof
public character,bestowedby membersof the municipal upper class.96
This sameclasspresumablyintroducedthe luxuries of the town housestoo.
The mosaicfloors in the settlementson the Danubeoriginate, practically with-
out exception,from the time of the Severi; the exceptions,like the mosaicsof
the governor'spalacein Aquincum (probably of mid-second-centurydate) be-
long to a socially quite different category.97It is not impossiblethat demandin
somelarge centresof the frontier areacausedthe appearanceof workshopsfor
the productionof mosaics.In Aquincum a workshopproducedseveralmosaic
floors with a compositionof large figures, rathercrudely executed(PI. 28b); the
compositionsgo back to known iconographicoriginals in Hellenistic art.98 But
also at Aquincum a mosaic floor cameto light recently,99of which the central
picture, showing the scenewith Hercules,Deianeiraand the centaurNessus
(PI. 28a), was certainly imported and insertedinto the floor made by the local
masters.Wall-paintings,occasionallywith large figural compositionstoo, were
also in demand,and walls and ceilings were often adornedwith stucco.IOO To all
appearancethere were no differencesbetweenthe two parts of the Danubian
towns: the luxury housesare found equally in what had beenthe municipium
and in the former canabae.
The luxury of the upper class showedin burials as well: in the whole of the
Danubianareaburial in sarcophagibecamefashionabletowards the end of the
secondcentury (PI. 29); in some towns such as Viminacium it was possibleto
establishthat from the end of the second century onwards rich citizens no
longer used tombstonesbut greatergrave-monumentsor subterraneanburial-
chambers,in which the often richly decoratedstonesarcophagiwere placed.lor
This fashion appearedearliest perhapsat Ratiaria where several sarcophagi,
237
possibly going back to the middle of the secondcentury,are derivedfrom gar-
land sarcophagiof Asia Minor. l02 The spreadof this customalong the Danube
caneasily be followed.I03 The changeto burial in sarcophagimeant,of course,at
the sametime a changeto inhumationand abandonmentof the rite of cremation.
This changein rite took place first among the rich inhabitantsof the towns;
whether it originated from the Orientals, as has been frequently maintained,
cannot be decided.It was at any rate almost invariable that the richest people
in the towns on the Danubein the third century did not set up tombstonesbut
sarcophagior great funerary monuments.The fact that many Orientals were
among them will not be surprising in view of the part played by this ethnic
component,which hasalreadybeendescribed.
In the period of the Severi the decurions of the Danubian towns to all
appearancelived inside the towns where they displayedtheir building activity
and wherethey were buried. In the rural surroundingsof thesetowns no great
villas that could be interpretedas the centresof larger estatesand as seatsof
estate-owners areso far attested.The only exceptionseemsto be the areasouthof
Carnuntum,where the estatesof the Boian aristocracylay. A few smallervillas
of this region which have beeninvestigatedwere abandoned,accordingto the
stratigraphicalevidence,in the secondhalf of the secondcentury,I04presumably
as the result of the estatesbeing incorporatedin larger estates.However, the
largestof the villas, the palaceof Parndorf(Fig. 32), seemsto have beenlived
in and usedwithout interruption. It was the centreof a presumablyvery large
estateor evena latifundium. In the remainingparts of the frontier zoneno large
estatescan be demonstrated,and this is the reasonfor our venturingto draw the
conclusion,already hinted at above, that the wealth of the upper class in this
areawas not basedon the concentrationof agrarianestates.
The existenceof a well-to-do class in the country cannot, nevertheless,be
denied. Apart from monuments,such as the huge burial-vault from Brestovik
(Pl. 30), eastof Singidunum,which was built on the model of grave-chambers
in ThraceandAsia Minor andadornedwith wall-paintings,andwhich was at the
time apparentlyconsideredexceptional,105 all settlementsand cemeteriesun-
coveredin the border areaso far suggesteconomicprosperityand the spreadof
the Romanway of life. In the civil settlementsof the auxiliary forts of Pannonia
it canbe observedthat with few exceptionsstonehousesdo not occurbeforethe
time of the Severi. This is the casenot only at Intercisa,where the Syrian en-
clave createda situation which must not be generalized,but also at those of
Ulcisia Castra,Albertfalva and Matrica (Fig. 54), of which the civil settlements
have been investigatedto a certain extent.l06 The same is true of the non-
military settlementsof the frontier area and of the hinterland of the limes. As
23 8
has been statedin Chapter 5, no villas of early imperial date-norany stone
houses-areso far known in north or easternPannonia.107 In the surroundings
of Aquincum,however,whereexcavationshavebeencarriedout in manyplaces,
smallervillas and stonehousescould be examined;their constructionhas been
dated,or can be dated,to the Severanperiod. Thesehousesare heatedby hypo-
causts,and have paintedwalls and the like too, but their size is mostly smaller
than that of the villas of early imperial datein the westernand interior parts of
Pannonia.Thesevillas are particularly frequentin the immediatesurroundings
of Aquincum.108 A group of fairly humblehousesnot far from Aquincum had a
cemeteryof its own in which the grave-inscriptionsof the inhabitantscameto
light.r 09 From the nameof a woman,Bithynia Severa,it canbe inferred that the
owners were Orientals. As we have seen,there are some indications that the
wealth of the upper class in the frontier areadid not derive from agricultural
estates.The origin of its prosperitymust be lookedfor elsewhere.Our starting-
point must be the circumstancewhich is perhapsmost characteristicof society
in the frontier area,that no strict distinctioncan bemadebetweenthe municipal
(civil) upper classand the military. One of the reasonswas that decurionsand
garrisonswere closely interwoven: there are caseswhere the sonof a decurion
enteredthe legion,l 10 and in some towns the veteransformed a continual re-
plenishmentfor the ordo of the town.III Many private soldiersseemto havebeen
no less wealthy than the prosperouscivilians. We havefor examplea fair num-
ber of sarcophagiset up by active or dischargedsoldiersof the legions.I12 The
great number of family gravestonesbelonging to legionariesnot only attests
regular recruitmentfrom the immediate neighbourhoodof the garrison but,
evenmore, closeinterconnectionbetweenthe local populationand the military.
The village inhabitantsprobablyprovided recruits readily, and military service
yielded a regular and not negligible income for small family units. Apparently
this was the reasonwhy the formation of large estatesdid not takeplacein the
borderarea.The small estatesof the village populationin the frontier zonewere
able to competeand could hold their own becausethey could count on a regular
income from outside whichcame to them throughthe military. The 'normal'
developmentin the direction of town-basedestatesand finally of large estates-
these latter already outside the limits of municipal economy-wasthus im-
pededin the border area,and remainedso for as long as the military had at its
disposalan incomefrom outsidethe economyof its own province.
This hypothesisfinds its confirmation in a comparisonbetweenthe circum-
stancesof the borderareaand thoseof the interior of the province.As we have
seen,a concentrationof estatesin the interior of Pannoniawas alreadygoing on
during the secondcentury.A comparisonof thedistributionof tomb-inscriptions
239
in the ordinary countrysidewith those areas where villas and-by means of
villas and otherindications-largeestatescan be assumedto haveexistedshows
immediately that thesetwo distributions are more or less mutually exclusive.
Wherevillas or largeestatescanbe demonstrated,tomb-inscriptionsarelacking,
and vice versa. Since grave-inscriptionsin the open country mostly represent
families which possessedat the most only moderateprosperityand were very
often closely connectedwith the legions inthe vicinity, we can distinguishtwo
types of social and economic development. These are the aforementioned
'normal'developmenttowards the concentrationof estateson the onehandand
on the other, developmentassociatedwith the prosperityof small estatesin the
borderzonewhich were closelyconnectedwith the military. The place-namesof
inner Pannoniafurnish further proofs: there is a small numberof place-names
formed from a native nameor from a cognomen,such as Mariniana, Maestriana,
Bassiana.Thesenamescan be consideredas adjectivesto the word villa, e.g.
villa Mariniana = 'Villa of Marinus', and thus may go back to villa establish-
ments of early or middle imperial date. Their distribution is confined to the
areasof 'normal' development(Fig. 38, p. 233).
Socialdevelopmentin Upper Moesia in the third century

The interior of Pannoniabelongs as a whole to the zone of normal devloep-
ment. In Upper Moesia, the situation was probably altogetherdifferent. We
must first notice that the mines formed a factor not presentin Pannonia,and
secondly, that recruitment was late in starting throughout the whole of this
province, while the areaswhich provided most of the recruits (Scupi, Ratiaria)
did not lie in the bordersection-whichwas anyhowshort. Moreover,the army
of Upper Moesia drew to a much larger extent on the population of the un-
armed adjacentprovinces. For this reasonand also becausetension between
original inhabitantsand aliens was alwaysperceptible,the military could not act
as a connectinglink betweenthe provincial upper class and the native village
population.Apart from ScupiandUlpianum,the towns of the interior ofMoesia
were all creationsof Marcus and the Severi. They were never really able to de-
velop a municipallife of their own, despitethe obvioustendenciesof the govern-
ment to include the local upper class in the municipal administration.In the
earliertowns, managementhadbeenin the handsof foreign settlersandbusiness-
menwho wereratherisolatedfrom the nativepopulation.Ulpianum,whosearis-
tocracy-uniquein the province-consisted, in the secondcentury,of Dardani-
ans, had the normal development,as is shown by the absenceof tombstonesof
military families; in the third century it did, in fact, provide recruits,!I3 It is
240
precisely in the territorium of Ulpianum that one of the few latifundia of the
Danubianareacan be demonstrated:this was the great estateof the senatorial
families of the Furii and Pontii on the northernslope of the ScardusMons, the
slave personnelof which left behind quite a numberof inscriptionsin the Sev-
eran period.II 4 A freedmanof these senators,Furius Alcimus, is known, too,
as the recipient of a responsumfrom the jurist Ulpian.IIs
Except at the two legionary fortresses,tombstonesof soldiers' families are
attestedin Upper Moesia only at Timacum minus, where lay the fort of cohors
II Aurelia Dardanorumand in the valley of the Morava (Horreum Margi and
Naissus).II6The families of soldiersat Singidunum,Viminacium and Timacum
minus were of mixed foreign provenance,continually settling nearthe garrison
and forming a hereditarymilitary profession.The sons of soldiersin Timacum
minus frequently enteredthe legion even as early as the secondcentury. These
families had hadpracticallyno connectionat all with the local populationof the
district. The fact that from the end of the secondcentury onwards recruiting
for the legions of Upper Moesia was carried on outsidethe military coloniae of
Scupi and Ratiaria is proved, apart from somestatementsof origin by soldiers
from the city of Rome, only by family tombstonesfrom Horreum Margi and
from the territory of Naissus. It might be assumedthat the formation of a
social class consistingof small peasantsand soldiers-afterthe patternof the
Pannonianfrontier districts-took place in this area, but suspicionis roused
by the circumstancethat the prosperity attested by these tombstoneshad
come into being along the very road that played such an important role in
the third century. The causesof prosperity were thus not necessarilyof a
politico-military kind. The seedsof a developingprosperitycannot,however,
be denied in some parts of Upper Moesia. The difference betweenPannonia
and Upper Moesia is probably based on the fact that, owing to the late be-
ginning of local recruiting, a population of military complexion was un-
able to come into being until very late, if at all. The military of this province
was composed,much more than that of Pannonia,of a foreign class of here-
ditary soldiers, who were isolated from the population of the province. The
causeshad deep local roots which were discussedin Chapter5 :for reasonsof
security local recruiting could not be spreadover the whole of the province,
only few Romanself-governingcities were founded before Severantimes, and
the original populationwas thus excludedfrom the leadershipof the local life of
the community.
241
The Il(yriciani
What has been said gives us some clues to the social characteristicsof the
Danubearmy at the culminationof its political activity, as well as to its historical
role. A first, very typical componentwas formed by the residentfamilies in the
villages of the frontier area,who had an economicconnectionwith the provin-
cial army, in that the income of their members,serving in neighbouringunits,
renderedthe family estatescompetitive.This social class was neither a peasant
soldiery nor a hereditarymilitary cast. It had the opportunityof rising into the
higher military ranks as well as into the civilian urban upperclass, which was
not strictly isolatedfrom it. It was, however, continually reinforcedfrom the
circles of the local native population. A retrospectivecharacterizationof the
Illyriciani in Aurelius Victor apparentlyhad this peasantclassin mind: his sane
omnibusII(yricum patria juit, qui quamquamhumanitatisparum,ruris tamenac militiae
miseriis imbuti satis optimi rei publicae juere ('Almost all these were natives of
Illyricum; though lacking culture they were rearedon the hardshipsof rural
and military life and were thereforegood servantsof the state.')II7
A secondcomponentwas the military families, long residentin the vicinity
of the garrisons;with them military servicehad practically becomehereditary.
Thesefamilies presumablyprovided most of the centurionsand higher posts
in the army and provincial administrations,for the officers of the Danubearmy
in the third centurystill seemto have beenmostly peoplewith Italian nominaor
with older imperial family nameswho cannothavecomefrom the rural popula-
tion of the frontier region.II8 Reinforcementfor this hereditaryprofessionof
soldiers was ensuredby recruits from the unarmedDanubianand Balkan pro-
vinces (Dalmatia, Thrace and Macedonia), and from the easternprovinces.
Recruits from those interior regions of military provinces which were not
occupiedby the army must be reckonedas belongingheretoo, since it may be
consideredunlikely that the inhabitantsof areasof 'normal' developmentsought
to join the army unlessthey were poor or had lost their estates.For the inhabi-
tants of the frontier area,however,serviceas soldierswas a meansof maintain-
ing the prosperity of their estatesand other undertakings.The non-agrarian
enterprisesin the border area in particular were directly dependenton the
army, as was the economicand communitylife of the borderareaas a whole.
F or the samereasonthe civil populationof the towns in the frontier areawas
closely dependenton the army, too. As we have shown,large estateswere un-
able to developjust becausethe rural small estatescould draw from the sources
of the military economy. The same was presumablytrue of commercial and
industrial undertakings,since it is a significant characteristicof economiclife
242
in the frontier provincesthat hardly any industrial centrescapableof achieving
somethingfor the economyof the empirearosein them. We shall return to this
problemlater in this chapter.Tradewas thus chiefly an import- or transit-trade
acrossthe frontier, and economiclife was basedmuch more on consumption
than on production. Under such circumstancesa social contrastbetweenrural
soldiery and urban citizens, such as was imaginedfor instanceby Rostovtzeff,
could not arise, at least not in those areaswhere the army was stationed-or
rather, not only stationedbut so intertwined with local society as to give this
society its most significant characteristic.
A certain democratictendencyin the aspirationsof the Illyriciani is, never-
theless,undeniable.The soldiersof the Danubearmy were in origin peasants-
indeed,peasantsbelongingto an agricultural structurethat knew neither slave
labour nor coloni. The most influential social elementwas presumablyformed
by them,not by the soldiersfrom the hereditarymilitary families in the garrison-
towns, whosesole basisof existencewas their affiliation with the army. As pro-
fessionalsoldierstheir sole interestwas presumablythat their existenceshould
be securedby the imperial government.Thus they were excellenttools in the
hands of any military movement that wished to count on their help. It is
thereforejustifiable to speakof an Illyrican soldiery.
Unfortunatelythe percentagefigures of the two componentelementsof the
Danubearmy are unknown. The shareof the professionalsoldierspresumably
varieda little from legion to legion; it was,for instance,presumablymuchlarger
in legio VII Claudiathan in the legions of PannoniaInferior (I and II Adiutrix),
whererural soldiers'families areattestedin the borderareaby manytombstones.
The participationof the professionalsoldiers must also have beenlarge in the
Upper Pannonianlegions X and XIV Gemina, since large estates,developed
from the Boian estates,lay in the very hinterlandof Carnuntum;and elsewhere
in this province,too, there are a numberof proofs of 'normal' development.
Economicdevelopmentsoj the third century

The Danuberegion's economicand political centre of gravity lay in the fron-
tier zone.A certainwealth andprosperityin the interior of the provincescannot
be denied,as we have alreadyseen,but strongersocial polarizationthere set a
barrier to this prosperity. This polarizationis attested,above all, by the large
estates,the full developmentof which falls into the Severanperiod. Thosevilla
establishmentsin the interior of Pannoniawhich had originated in an earlier
period continuedin use, and some of them were very luxuriously appointed.
Most famousarethe largemosaicfloors from Balaca,north of Lake Pelso,which
243
haverecentlybeendatedto late or post-Severantimes.119 It is very probablethat,
as a result of the concentrationof estates,somesmallervillas were abandoned;
this could be demonstratedarchaeologicallyat severalsites in the territory of
the Boii near Lake Neusiedl (Lake Ferta in Hungarian).The wider surround-
ings of Sirmium, that is the country betweenDanubeand Save,must havebeen
a region of large estatestoo. The SenatorC. MessiusDecius,proclaimedby the
Illyriciani as the first Pannonianemperor,was born in Budalia, a village west of
Sirmium, on the Sirmium-Cibalaeroad,120 presumablyin a family villa. The
EmperorProbus,too, had estatesaround Sirmium:121 it was on their account
that he was killed by his soldiers.In this fertile countryaroundSirmium objects
of luxury and sculpturesof good quality have come to light as unassociated
finds in large numbers;they allow us to infer richly appointedrural seats.l22
Possibly during the third century some of the large estatespassedinto im-
perial possession.Some imperial estatesoccur even earlier, for instance one
somewherenearRatiaria,123andanothernot far from Sirmium124 whereHerodes
Atticus stayedin 174. The place-nameCaesarianaon the Savaria-Aquincum
road, in the region of large estatesnorth-eastof Lake Pelso, also points to an
imperial estate.It can be assumedat once that during the courseof the third
centurya numberof large estatespassedinto imperial possessionas membersof
the aristocracywere killed or exiled in the courseof political upheavals.The
family estateof the Emperor Decius apparentlybecamesuch a domain, but a
changein private ownership in the third century has to be assumedfor the
palace-likevilla of Parndorfin the country of the Boii. From the beginningof
the fourth century onwards we have an imposing number of demonstrably
imperial domains,in Pannoniaas well as in Upper Moesia.
The owners of large estatesin the interior of Pannoniawere descendantsof
the municipal aristocracyof early imperial times. A certainregressionof munici-
pal life, shown in the small number of building- and votive inscriptions of
decurions,canpresumablybe attributedto increasingwithdrawal by prosperous
membersof the ordo from municipal affairs. Of course,new buildings are not
wanting in the towns evenafter Marcus;for instancein Savariaa greatshrineof
Isis was built, possibly on the site of an earlier temple (Fig. 41, PI. 27b).125 In
Savaria,too, mosaicsof the Severanperiodareattested.The prosperousmiddle
classesof the populationof both town and country, however,havedisappeared.
The tombstoneswhich are characteristicof this social classbecomerarer evenin
those placeswhere they had beenexceedinglyfrequent in the first and second
centuries; as at Savaria,Poetovio and in the region of Lake Pelso. Imperial
endowmentsand other measures,like the Kalendarium Septimianumin Savaria
(p. 219) or the colonia (nova) in Siscia, under such circumstancespresumably
244
The secondage oj prosperity.. rise and collapse
servedthe purposeof preventingregressionof municipal life. Theseattempts,
however,have left no trace of any notablesuccess.One of the reasonsfor this
lack of successmay have beenthe ability of wealthy membersof the municipal
aristocracyto use the prosperity of the times to further their own interests;
they did not representa social stratumwhich for political reasonscould be, or
had to be, treatedto its disadvantage.As early as the secondcenturythis class
had given capableofficers to the empire, like the famous M. Valerius Maxi-
mianusfrom Poetovio,and a numberof equestrians,someof whom are attested
in various positionsof the tres militiae. 126 Though the notoriousself-mademen
of the Severanperiod mainly originated from military circles in the frontier
area,like Aelius Triccianus,who startedhis careerin the officium of the gov-
ernor,127later representativesof the Illyriciani were Sirmians and so presum-
ably men of wealth, or else the majority of them possessed family nameswhich
allow us to infer a citizenshipof older origin; they cannotthereforebe takento
representthe original inhabitantsof the frontier area.128 Decius had attempted
to avoid supportingthe causeof the Illyriciani, but the confidencewhich the
latter had in him proves that a conflict of interests between the military
and the prosperousupper classesof the areasof 'normal development'did not
exist.
No considerablechangesin the trade-relationsof Pannoniaand in the focal
points of its economiclife took placeat the end of the secondcentury.Imports,
demonstratedby finds, mostly comefrom the westernprovinces,with the differ-
ence that the areason the upper Danube,being situatednearerto Pannonia,
come more and more into the foregroundas suppliers.After Marcus, finds of
samian ware are almost exclusively Raetian (Westerndorfware).129 So-called
Raetianpottery is also well representedin Pannonia,but the majority of the
productswas presumablymanufacturedon the spot.130The sameshouldbe true
of the glassware,which, though in shapebasedon products of the Rhenish
factories,was for the most partproducedin Pannonia.I3I Imports from the west
included enamelled bronze brooches and other types of enamelled bronze
ornaments;thesewere also imported by Pannoniafor the transit-trade:enam-
elled bronzes were very popular among the Sarmatiansof the Hungarian
plain.I32 Rhenishmotto-beakers,bronzemounts,carvingsin jet, etc., are found
on all the larger sites.
Importation in equally large quantitiesfrom the easternprovincescannotbe
proved. A few striking objects do however come from workshopsin Syria,
Egypt or Asia Minor. Without exceptionthey are luxury goodsor small works
of art suchas bronzestatuettes,133sculpturesin ivory 134 and richly ornamented
silver or bronze vessels135 producedin the workshopsof Alexandria. A con-
245
siderablenumber of pieces of gold jewelry presumablyalso comes from the
East, as well as the glass-pasteimitations of cut stones,often showing Greek
mottoes.136 Fine textiles were apparentlybrought from the Orient; they can
only rarely be tracedamongour finds.137
At this period oriental imports presumablyno longer came to Pannoniaby
way of Aquileia, but already on the Byzantium-Naissus-Singidunum road
throughthe Balkansor to an evengreaterextentby water, on the Danube.And
becauseof this trade-routeon the river we may assumethat orientalwaresplayed
an even greaterpart in Upper Moesia'simport trade. Becausework on small
finds has not progressedvery far proofs are not yet available,but oriental in-
fluences show fairly directly in stone sculptureand they may even perhapsbe
representedby imported piecessuch as the garland sarcophagiat Ratiaria and
possibly at Viminacium too.I38
In general,finds of Severandate are much more uniform, much less various,
than thoseof earlier times, the chief reasonbeing that the local productionof
everydayobjects managedto displaceimports almosteverywhere,evenwhere
the quality of the local goods was inferior to that of the imports. Imported
goodswere soonimitatedon the spot. It hasfor instanceonly recentlybeendis-
coveredthat so-called Raetian pottery was also made in Pannonia.I39 In the
sameway, it becomesincreasingly probable that most of the glasswarewas
produced in Pannonian workshops. From not uncommon discoveries of
moulds for casting knee-brooches,buckles, mounts and similar objects140 it is
believedthat the simple bronzeobjectsof usein all centresof the provincewere
locally produced.
Nevertheless,industry in Pannoniaand Upper Moesia did not produceany
specialityof significancein the tradeof the empire. So far, researchhasnot suc-
ceededin demonstratingthat objects of Pannonianor Upper Moesian origin
exist in large numbers in other parts of the empire. Identifiably Pannonian
products were of hardly any significanceeconomically,and it cannoteven be
shown that they were distributedby the normal processesof trade; they could
just as well have owed their dispersalto troop-movements.ThesePannonian
productswere small votive tablets of the so-calledDanubianRider-god or of
Mithras, madefrom the soft yellowish crystallinelimestoneof the Alma Mons
(Fruska Gora).I41 Writers of the early imperial period inform us, furthermore,
on a numberof other Pannonianexports,suchas medicinalherbs,wild animals
for the amphitheatre,huntingdogsandpossiblyalsohorses.142 In their mountain
farms the Dardaniansof SouthMoesiaproduceda cheesespeciality.143It cannot
be said of coursehow theseexportsaffectedthe trade-balanceof our provinces.
Of greatestimportance,presumably,was the export of silver, lead and copper
246
from UpperMoesia;but as hasalreadybeenseen,it contributedto the prosperity
of only a limited numberof towns and settlements.
In the time of the Severiexporttradeacrossthe frontier was to all appearance
very significant. This exportis confirmedby indisputablyRomangoodsamong
the finds from sitesoutsidethe empireas well as by the influenceexertedby the
provincial industry on local barbarianpottery production.Barbarian-Roman
trade-relationsare demonstrated,too, by severaltypical kinds of objectsfound
in settlementsoutside forts along the Danube,for instancespecial shapesof
vesselsof Sarmatiantype.I44 For the centres onthe bank of the Danubethis
trade presumablyformed a componentof prosperity, but from an imperial
view-point it may havebroughtbut little gain, especiallysincethe treatieswith
barbarianborder peoplesarrangedfor their regular subsidy. Roman exports
thus tendedto be boughtby barbarianswith Romanmoney. No evidenceat all
existsfor any importantimport-tradefrom the barbariansinto Upper Moesiaor
Pannonia.I45 The presumablybusy trade acrossthe border along the Danube,
though a local sourceof gain for the borderareas,was quite expensivefor the
empire. Therefore the encouragementprovided by the border trade for the
economyof the frontier region must be countedamongthe phenomenaof the
period of prosperity.
Standardizationoj local culture in the third century

Uniformity of industrial production went hand in hand with a standardiza-
tion of local culture. In all aspectsof culture there becomesevident a fading
of the local characteristicswhich in the secondcenturyin particularhad shown
up very significantly. This standardizationwas so intensiveas to causethe al-
most completedisappearance of local, non-Romannomenclaturein Pannonia;
specifically local rites of burial (tumuli, wagon-graves),local burial symbolism
(astral symbols)and much else besidesdisappearedas well. This comparatively
swift processwhich took placeat the endof the secondcenturyhasbeenheld to
meanthat the native populationhad beenso reducedby the Marcomannicwars
as to render necessarythe oriental and other immigrations for the purposeof
populatingthe country again. But the war under Marcus,althoughundeniably
catastrophic,did not causesuch an ethnic vacuum. Celtic namessporadically
continueto occur after Marcus, particularly in those regionswhere epigraphic
evidencefor local nomenclaturein the first 'and secondcenturiesis densest.
Thesesporadicnames,now appearingas cognominain the three-namessystem,
are carried by personswhose families were already changing to the Roman
systemof nomenclature. 146 This Roman nomenclatureis very colourless:the
247
only feature which is to someextent characteristicis the preferencefor certain
cognominasuch as Decoratus,Candidus,etc. A few further namesare successful
Latinizationsof Celtic or Illyrian ones,suchas the cognomenof Aelius Triccianus,
formed from the Celtic Triccus.
In SeveranPannoniathe widest retention of non-Latin nomenclaturepre-
sumably occurred among the barbarian groups settled under Marcus in the
province. But thesegroupshaveleft behindhardly any epigraphicmonuments.
A group of Cotini were settled under Marcus in the territories of Mursa and
Cibalae;alreadyunderthe Severithey were enteringthe legions,and someeven
rose to positions in the praetorian guard. In Rome, these are listed in the
laterct/li Praetorianorttm, with cognominathat suggestan intenseDacianizationof
the Cotini.I47
EnergeticLatinization of nomenclaturein Pannoniacan,of course,be proved
only with referenceto the elementsnamedon inscriptions. However, the fact
that this 'epigraphicallytraceable'social classall of a suddengaveup their non-
Latin nomenclature,to which as far as we can tell they had clung tenaciously
during the secondcentury, is a symptomaticand important phenomenon.As
we haveseenin Chapter5, theseepigraphicallytraceableCelts andIllyrians were
the chief componentsof the local aristocracywhich enjoyedthe favour of the
government,and they retained certain traditions of their culture despite or
becauseof their Romanization.If this attachmentto tradition on the part of
the local upper class was a consciousattitude in the secondcentury, then the
abandonmentof this tradition must be considered as equally conscious.
Their increasingshareof local authority, evenmore their growing political im-
portance in the empire, apparentlyrenderedan adaptationto Roman forms
necessary.
The relinquishmentof tumuli and wagon-burialsprobablyalso belongsin this
context. Hitherto, no tumuli indisputablydatableto the third centuryhave been
found. A few cart-gravesmay possibly dateto the early Severanperiod, but in
the third century this expensiverite fell completely into disuse, and certainly
long before the economic crisis reached the Danube; the disappearanceof
the fashion cannot therefore be attributed to economic reasons. Among
typical local representationsof tombstonesymbolism, astral symbols and re-
presentationsof chariots, too, disappearedsoon after the end of the second
century;the only iconographiccompositionto continuein usewasthe onewhich
was leastun-Roman:this was the so-calledsceneof sacrificewith the tripod and
sacrificing figures standingon either side (camilli). In the period of the Severi
this sceneis found on the tombstonesof non-local families too.I48
Among the monumentsof Upper Moesia, such a recessionof native tradi-
248
The secondage oj prosperity,' rise and collapse
tions cannot be traced, mainly for the reasonthat the native population as a
result of insufficient adaptationto Roman influence had never hitherto pro-
duced anything specific. It is only from the beginning of the Severanperiod
that we can speakof adjustmentwhich affectedthe whole province.It was then
that most of the self-governingtowns were founded. The consequences of this
late consolidationbecameapparentonly towards the middle of the third cen-
tury. With the exceptionof the inhabitantsof the territorium of Ulpianum, the
original populationof Upper Moesiaonly very rarely set up inscriptions:it was
done by a few soldiers'families in the valley of the Morava and somevery few
decurionsof the towns of late foundation in the interior of the province. This
epigraphicallytraceablestratumof the native population,however,alreadycar-
ried Latin names.Particularly instructive in this respectis the extensivelist of
169 recruited and 195 dischargedsoldiers of legio VII Claudia;I49 it enumer-
ates one age-groupwithout gapsand thus renderspossiblea comparisonwith
the compositionof the epigraphicallytraceablelegionaries.On this list, a fairly
large number of Upper Moesian soldiers with Thracian cognominaare named
(from time to time Illyrian cognominaoccur, too), whereason the other inscrip-
tions of the province legionaries with Thracian (or Illyrian) cognomina are
excessivelyrare. From this fact it is possibleto concludethat a large mass of
alreadyfairly Romanizednativesare not traceableon inscriptionssuchas grave-
inscriptions and altars. Only after the middle of the third century do they be-
come traceablehere and there, and on inscriptions they often appear with
Thracian names-unlikecontemporaryPannonians.A certain Thracian self-
assurance,too, is seenwhen a town-born Roman soldier, a citizen of Scupi,
gives his origin as Bessus.ISO This indication of origin has causedconsiderable
difficulty, sincethe clan of the Bessi were not settledin the territorium of Scupi,
nor evenin Upper Moesiaat all. In late usage,however,Bessussimply signifies
someonespeaking Thracian; presumablyby the name Bessusthe soldier of
Scupi merely expressedhis membershipof the Thracian-speaking inhabitantsof
the empire.
It cannotbe deniedthat this developingself-assurance is somehowconnected
with the self-assurance of the Illyriciani. Particularisticself-assurance
andempha-
sis on genuineRomanity were equally characteristicof thesesoldiers,and their
Danubianorigin often gavean opportunityto show genuineold-Romantraits.
Thus his origin in Dardania becamea proof that the Dardanianin question
(the Emperor ClaudiusII) was in the last resort a Trojan-in other words an
original Romanof the best kind.
249
Religiouscults in the third century

When the Pannonianssuddenly steppedon the stage of history in 193, and
put their official cult to the service of the Severanparty, they did not intend
to enlargetheir local pantheonand its local names,but chose Romandeities,
or at least laid emphasison deities and cults that could join the Roman cult
without difficulty. Pannonianaugurs are reported, who possibly go back to
Celtic usages,but who were as augursneverthelessacceptableto any Roman;151
the Matres Pannoniorum,too, could easily be understoodby all, althoughthey
may possibly descendfrom the Nutrices of Poetovio.I52 Later on in Pannonia
itself cults madetheir appearancewhich, although they may be taken to repre-
sent real Pannoniancults, do not show truly local, non-Romantraits. Possibly
the most essentialtrait of thesecults was that everythingaboutthem-thename
of the deity, appearance,iconography-wasRoman, or at least, they did not
show any local traits not elsewhereattestedin the empire. This particularity of
the cults and of the religion of Pannoniacan be understoodonly when seenin
connectionwith the process we have just described: the suddenfading of
local, native peculiaritiesof the culture in favour of a standardized,colourless
Roman attitude, which, in the last resort, was based on Danubian-Roman
self-assurance.
In the cult of SilvanusI 53 in particular quite a number of difficulties are re-
moved if it is interpretedin this connection.These difficulties had arisen be-
causethose investigatingthe cult wished simply to seein Silvanus an Illyrian
deity, if not the chief god of the Illyrians. This conclusionreachedlong ago was
basedmore or lessonly on the circumstancethat an exceedinglylargenumberof
monumentsin Pannoniawas dedicatedto the god Silvanus;in somesettlements
the number of altars to Silvanus exceeds eventhose to Juppiter Optimus
Maximus; and in generalSilvanusis the god with the greatestnumberof altars
after Juppiter. The floruit of the Silvanus-cultfalls into the Severanperiod.
(The earlier monumentsof the cult are found only in the towns of the Amber
Route and have connectionswith the cult of SilvanusAugustusin Aquileia,I54
who cannot be consideredas a native god.) A further point in favour of an
interpretationof Silvanusas a renamedIllyrian or, more generally,native Pan-
nonian deity, is the connectionof some aspectsof his cult with pre-Roman
fertility-cults, especially with ithyphallic representationsof Iron Age date.I55
No doubt Silvanus is from time to time representedin an ithyphallic style on
Pannonianreliefs too,l56 and the fertility conceptwas not alien to the Silvanus-
cult as a whole. Thesetraits are, however,too generallyRomanto be takenas a
local specializedtrait. As was pointed out severaldecadesago157 the Pannonian
25°
Silvanusshowsno traits of characterthat are not attestedin the Italian cult of
Silvanus also. It is thereforequite probablethat for their favourite native god
the Pannonianshad selectedan Italian deity; the fact that this god was not
simply basedon a pre-Romandeity of like naturefollows from theintroduction
of his worship only from the third century onwards,and then only in an un-
specificform.
It is not to be deniedthat in PannoniaSilvanuswas the most popular, most
frequently worshippeddeity in the period of the Severi (PI. 31a). He is repre-
sentedonly very rarely on altarswith othergods;he is mostly representedalone
or in companywith other deities of nature,relatedto him in character,like the
very colourlessSilvanae(occasionallynamedQuadriviaeand, when in the sin-
gular, Diana). With the gods of the classicalpantheonSilvanusis not brought
into connection,and even less with oriental mystery religions, though these
during the time of the Severi were widespreadtoo. On somealtars Silvanusis
actuallynamedin companywith the CapitolineTriad or with JuppiterOptimus
Maximus,I58 which bestowsa certain official characteron him. It is not im-
possible that as early as the beginning of the second century Silvanus was
officially depictedas the symbolof Pannonia.159 But if we takeinto consideration
the fact that a god namedSilvanusAugustuswas worshippedin Pannoniaat an
early date,then the conclusionsuggestsitself that Pannonia'sfigure of Silvanus
was in fact derived from outside. This is not at all surprising since elsewhere
too the interpretatio Romanaarosefrom the Romansputting somenative cults,
which they worshippedasgenii loci, on a par with their own gods.Silvanus,how-
ever,is not an exampleof an interpretedlocal god, or at leastnot of a singlegod,
since he has no known correspondingdeity with a native name. He has a few
non-Latin titles; 160 but each of these is only attestedonce, and already their
numbersallow the deductionthat one single Illyrian chief god, equatedwith
Silvanus, never existed. At one of the places dedicatedto Silvanus-it seems
that this god rarely had shrinesor temples-altarsfor an otherwiseunknown
god Vidasusand for a goddessThanacameto light. I61 This single casecannot
be taken,for the very reasonof its being unique,as an argumentin favour of an
interpretedIllyrian god.
The reasonswhy it was Silvanuswho becamethe most popular god of Pan-
nonia, and why this did not happenbefore Severantimes, can be most clearly
deducedfrom the characteristicsof the god himself. As we have alreadymade
clear, Silvanus did not becomethe god of a political idea. At the most he em-
bodiedthe provincePannoniain the symbolisminfluencedby literature,but the
Pannoniansthemselveswere not interestedin establishingSilvanusas the sym-
bol of their country. The most frequent cognomenof the god was the exact con-
25 1
trary of everythingofficial: Silvanusdomesticus,and most of the representations
are of this aspect.They show a simply dressedbeardedman with the knife of a
wine-groweror gardener,with fruit, with a tree and a watch- or sheep-dog. 162
In this contextthe cognomenherbarius also belongs,as do the associationof his
l63
worship with that of the Lares,164and the connectionswith fertility. All these
traits are taken from the Italian Silvanus, as is also the iconographictype of
Silvanus'representations. Silvanusdomesticusis the protectorof the garden,of the
estate,of the fruit andthe harvest(SilvanusMessor),and of fertility in general,as
far as agriculture,horticultureand cattle are concerned.As a god of decidedly
private characterhe is mostly worshippedon tiny altars in the houseand on
small reliefs, and often, too, on altars that do not statethe nameof the donor,
since they were meantfor domesticusage.I65 As a basisandprerequisitefor the
distribution of this conceptionof the deity the family estatesof small peasants
and the small propertiesof house,gardenand field apparentlyserved.166 As we
have alreadyexplainedin this chapter,the floruit of thesefamily-estatesin the
frontier areaexactly coincidedwith the Severanperiod, and the monumentsof
the cult of Silvanus,too, mostly originatefrom this area.The forms of the cult's
rites arepurely Roman:altarsfor daily domesticworshipanddomesticsacrifices;
small reliefs, also presumablyfor domesticshrinesor for the Lararia and so on.
Although the cult itself is closely connectedwith agriculture, it was hardly
practisedin the country; most of the monumentscomefrom the military settle-
ments on the Danube.
The remaining aspectsof Silvanus are Roman as well; best known was the
god Silvestris,a rural parallel for Domesticus.Silvestris often appearstogether
with thenymph-likeSilvanae,with Dianaandotherclassicalgoddesses of Nature;
67
and once he is explicitly consideredthe god of hunting.I This aspectof Sil-
vanus was mainly worshippedby the higher levels of local society and has a
certainliterary tinge-thatis to say it is providedwith classicaltraits and asso-
ciateddeities takenfrom the classicalpantheon.
That Silvanusis not an Illyrian god is conclusivelyshown by the fact of his
being worshippedmuch more in the northernfrontier zonethanin the Illyrian-
Pannoniansouth of the province.168 It is at any rate of the greatestsignificance
that the populargod of the Pannonians,andparticularlyof that part of the popu-
lation mostinfluencedby thearmy, was a god who in all his characteristicsandin
all the forms of his cult is purely Romanif not purely Italian. At the outsetthis
cult was presumablyspreadby the Italians,who were glad to discovertheir own
Silvanus-thetutor ftnium-in this land of small peasants,and in the end they
createda cult which was taken up by peasant soldiercircles.
The remaining gods, who are often traced back to native deities in the
25 2
specialiststudies,were also the products of introduction from abroad. It has
often been emphasizedthat the goddessDiana (PI. 3I b) was an Illyrian or at
leastnative deity (goddessof fertility and of the forest) in Romanattire; but it
should be rememberedthat in the literature of the early empire Pannoniaal-
readyfiguresasa denselywoodedcountryrich in game,169 andforeignersanxious
to placatethe divine powersof a strangeland would thereforenaturallyworship
Diana on the spot as a goddessparticularly presentin Pannonia.It would be
idle to enumerateall deities in whom local (Celtic andIllyrian) traits have been
discovered,for the most important circumstanceis that natives,and especially
those carrying local names,only exceptionallyset up altars or cult-reliefspO
Gods describedby their native, non-Latin names are to be consideredas
merelythe by-productsof a cult which in form andcontentswas entirelyRoman.
Genius CiniaemusI7I and Minitra 172 are probablythe only godsattestedso far
who are known nowhere but in Pannonia,and it is significant that they are
mentioned only once each. Sedatus,Juppiter Optimus Maximus Teutanus~
Epona, Mars Latobius, etc., who also enjoyed worship in Pannonia,I73were
introducedby foreigners,and they exhibit hardly anything suggestiveof local
roots. These divine figures, originating and given their characterelsewhere,
belong in Pannoniato the Roman pantheon, as do the Mystery religions,
practisedby town-dwellersandthe army. Our conceptionof cult andreligion in
Pannoniais controlledby what is preservedby meansof classicalRomanforms
-altars, reliefs and sanctuaries.The most characteristictrait of the religion of
Pannoniawas that it did not allow a synthesisof Romanand local conceptions
to arise; everythingwas expressedin Romantermsand the expressionitself was
not original. The fact that eventhis local religion doesnot becomereally trace-
able before the Severanperiod may be attributedin the last resort to the fact
that active participationin Roman civilization by the Pannonianarmy and its
closely linked civilian population only began to develop after the reign of
Marcus, as the result of the political activity of the Illyriciani.
This interpretationmakessenseof the even more colourlessreligion of the
provinceof UpperMoesia.The godsthemselvesandthe forms of cult do exhibit
a greatervariety, but only becauseUpperMoesialay on the borderbetweentwo
areasof culture: the province lay betweenthe Latin and the Greek worlds, or
more accuratelybetweenthe Latin-speakingworld of Italian and westernideas
and that of Thracianand oriental ideasexpressedin Greek. The 'native' gods of
Upper Moesia are in part purely Thracian deities whose names,forms of cult
and iconographywere developedin Thrace. They spreadin the districts ad-
jacentto that province. The otherinfluencecamefrom the north-west,bringing
southPannonian,Italian and generalRomancultS.I7 4 A Thraciangod of heaven,
253
weather,and of everythingsupernatural,seemsto have beenthe Juppiterwho
was worshippedunder the titles Culminalis, Epilophios, Paternus,Capitolinus
and so on.175 Epilophios being a Greek namepoints to a Greekformulation in
Thrace. None of these cults, however, won a distribution as wide as that of
Silvanusin Pannonia,althoughthey and presumablyseveralothers(Nemesisat
Ulpianum,I76Mars in the Timok valley,I77 etc.) arecharacteristicof somesmaller
areas,for instancethe title Paternusor Patronusin the north-westernparts,I78
and the Greekformulationsin the south-eastern.The only really local, specific-
ally UpperMoesian(Dardanian)god seemsto have beenAndinus,namedonce
only, on an altar dedicatedby a benejiciarius,I79i.e. by a foreigner. The names
Andinus, Andia, Andio, derived from that of the god, turn up quite often,
especiallyamongDardanians,!80
There is one particular form of cult with pronouncedcharacteristics,which
developedin Pannoniaor in UpperMoesia:this is the so-calledDanubianRider-
god (PI. 34a) or, more accurately,pair of Riders, after the patternof the Dios-
curi.I81 It is very likely that its fioruit fell within the later years of the third
century. His monumentsare small reliefs in lead or marble, with a complicated
crowd of different gods and symbols,composedin an order which is formalized
by tradition. The reliefs, of which severalhundredareknown, aredistributedin
the middle Balkanarea,DaciaandPannonia,but thecentreof the distributionlies
in southernPannoniaand northern Moesia, spreadingoutwards from these
areas.Nothing original was producedby this cult: all gods and symbols, and
even the iconographicsettingsare well-known types of classical,oriental and
other religions. An interpretationas a superstitiouspantheisticsynthesisof the
gods most worshippedtowardsthe middle of the third centuryis the best sug-
gestion;in this Sol, the gods of the stars,the Dioscuri, the symbolsof Mithras,
Epona, and much else were mingled; the only original feature is its strictly
organizedeclecticism.The representations on small reliefs may well be assigned
to the lowestlevels of religious expression;a few of themreacheddistantborder
areasin the empire, apparentlyamongthe belongingsof soldiers,I82I consider
it quitepossiblethatthis Rider-god,who mustnot be confusedwith the Thracian
Rider-hero, arosefrom pantheisticspeculationsin the circle of the Danubian
army; we must not forget that his centrewas the sameas that of the Illyriciani-
the surroundingsof Sirmium.
In this superstitiousandprimitive (becausetoo undifferentiatedand mechani-
cal) synthesisof theologicalideas, elementsof the Mystery religions were also
included. They constituted the most important religious movement in the
Danubianarea in Severantimes, though their diffusion had occurredearlier.
LeavingIsis aside,sinceher worshipin westernPannoniagoesbackto the north
254
CARNVNTVM
CARNVNTVM
CARNVNTVM
CARNVNTVM
second
second
second
CARNVNTVM
0 100 200
I km
Figure 40 Mithraea in Pannonia
Italians of the early period, we find that Mithras and Dolichenus made their
appearancein Pannoniaand Upper Moesia no later than the first half of the
secondcentury.Their first followers were soldiersfrom the eastandthe person-
nel of the publicum portorium Illyrici. The earliest temples of Mithras in both
Pannoniaand UpperMoesiawere built by the staff of the customs,by the slaves
of somestationesof the portorium Illyrici. I83 The distribution of the Mithraea so
far uncoveredor attestedby inscriptions suggeststhat the cult of Mithras was
spreadby regular missionaryactivity: 184 thesetemplesare attestedonly in cer-
tain towns and in their immediatesurroundings(Fig. 40). The fact that this mis-
sionaryactivity provedsuccessfulonly after the Marcomannicwar is presumably
explicable,not only in the presenceamongthe newly arrived Orientals of no
255
doubt numerousfollowers of oriental religions,18s but also in the changewhich
occurredtowardsthe endof the secondcenturyin the whole religious andspirit-
ual atmosphereof the Danubeprovinces.The abandonmentof local traditions
and the adaptationto generalRomanforms of civilization presumablyproved
propitious circumstancesfor the spreadof oriental religions, too.
Juppiter Dolichenus was worshippedhere as elsewheremainly in military
areas,although noteworthy monumentsof his cult have come from entirely
civilian settlements,for instanceSavaria.I86 The cult communitiesseemto have
beenwell organized:the priestsof Dolichenusthroughoutthe whole of Lower
Pannoniaassembledfor an unknown reasonand immortalizedthis event on a
votive altar.I87 Possiblythis occasionwas the visit by SeptimiusSeverusin 202,
for this emperorseemsto havebeena promoterof the cult. Whetherthe spread
of the cult among the Pannonianarmy was the result of Severanpolicies or
whether on the contrary Severusfavoured the cult becauseof its popularity
there, is not known; but it is worth noting that one of the earliest, if not the
earliest, monumentto Dolichenusin the westernprovinceswas found at Car-
nuntumwherethe first of the Severiwas proclaimedemperor.Two sanctuaries
of Dolichenusare known in Pannonia,at Carnuntumand at Brigetio (Fig. 41).
They were small rectangularrooms, strikingly narrow and modest, and they
stoodin the vicinity of other cult-roomsin an unpretentioustemple-quarter. 188
Recently a Dolichenumwith the whole of its cult-equipmentcameto light at

Egeta,a fort-site on the bank of the Danubein UpperMoesia.189 This sanctuary
was an oval building, small andlacking architecturalfeatures.It is strangethat it
also containeda dedicationto Mithras. The Dolichenum of Brigetio was ad-
jacentto a Mithraeum:thereareno indicationsthat the easternMystery religions
were mutually competitive.
It is unnecessaryto describethe numerousMithraea of Pannoniain detail
as they do not differ from the usualtype. The cult-relief of the chapelis a special
feature,however,for it belongsto a distinct Danubianiconographictype. The
most important centres of Mithraic religion, from which missionary activity
started(Fig. 40), were at Poetovio (possibly the earliest centre in Pannonia),
CarnuntumandAquincum.In thesetownsseveralMithraeahavebeenexcavated,
and they also occur in the wider surroundingsof thesetowns, whereasin some
of the other towns such chapelscan be demonstratedneither archaeologically
nor epigraphicallyand are absentin the neighbourhoodtoo. The Mithraic com-
munities mostly consistedof a group of people who were associatesin public
life: employeesof the customs'administration,family and domestic staff of a
wealthy man, soldiersin a regiment.Moreover, the topographicaldistribution
of Mithraea at Aquincum and Poetoviosuggeststhe mutual separationof their
25 6
SAVARIA TAC (GORS1VM)

ISEVM }{ TEMPlVM PROVINCIAE
}{
metreso,,",' __10'1:::=:=,,",'
15 20
_1liii'
Mlthraeum
Mlthraeum
Mithraeum Mlthraeum
Mlthraeum
We!'1
AQVINCVM
POETOVIO
D~OliChe.nl)m 0 Mlthraeum
Mlthraeum
=======:
BRIGETLO
: EGETA
o, }{ }{ }{
CARNVNTVM Dolichenum Dolichenum
Figure 4I Plans of templesand shrines
257
spheresof influenceratherin the mannerthat Christiancommunitiesweresepar-
ated. A few more centresof the cult must be supposedin the less-investigated
valley of the Save. The existenceof a Mithraeum in the surroundingsof Sir-
mium hasbeenprovedonly recently.I90The cult-centresin UpperMoesiacannot
at presentbe accuratelydetermined.The rather largernumber of monuments
in the DanubeValley round Singidunumand Viminacium point to Mithraic
communitiesin thesetowns.
The remaining oriental Mystery and religious movementsfound much less
responseon the Danube. The cult of Sarapishas a number of monumentsin
southernMoesia and Pannonia;191 its connectionwith the religious policies of
Caracallais sometimesvery probable.I92 The Egyptian cults to all appearance
survived only till the early years of the Severanperiod, but thesemarkedonly
the start of Mithraic prosperity.Thereare, of course,also, no small quantity of
monumentsrelating to cults from Syria and Asia Minor, but theseoften lack
clear definition in the Danube area: we cannot even be sure that they came
directly from the Orient. Dea Syria for instancepossesses a number of monu-
mentsin Brigetio, Aquincum and elsewhere,but her cult is missingin the very
settlementwhere theinhabitantshad most preservedtheir Syrian traditions, at
Intercisa. The distribution of all oriental religions in Pannoniaand Moesia
Superioris a phenomenonthat we must interpret in relation to the religious
history of the Romanempireas a whole ratherthan to the specialcircumstances
of theseprovincesthemselves(for instanceas the result of local oriental immi-
grations).Orientalsby no meansabsentedthemselvesfrom thesecults, but the
part they played in the mission was no greaterthan anywhereelse in the em-
pire.I93
Only the cult of Dolichenusseemsto be somehowclosely connectedwith the
Pannonianarmy, and can possibly therefore be related to particular traits of
Pannoniandevelopment.As alreadynoted, SeptimiusSeveruswas a promoter
of this cult. It is thereforepresumablynot coincidentalthat, of all the provinces
of the empire, Pannoniahas most monumentsof Dolichenus, and that the
spreadof the cult in the westernprovincesis often attributedto the missionary
activity of Pannoniansoldiersin particular.I 94 The puzzlingdisappearance of the
cult in the secondhalf of the third century is suspiciousenoughto suggest
political reasonsas its cause.To the Christiansat the beginning of the fourth
century Dolichenuswas no longer an enemyworth fighting. Thus the polemic
writings of theChristians,which enumerated all sortsof paganbeliefs andsuper-
stitions, do not mention Dolichenusat all. It is possibly rash to draw conclu-
sions from the circumstancesin which the Dolichenussanctuarieswere found,
but it is a remarkablefact that all their furnishingshavebeenpreserved,whereas
25 8
in the shrinesof Mithras, for instance,small finds belongingto the equipmentof
the cult are very rare. It seemsthereforethat the Dolichenussanctuarieswere
destroyedby a single violent act, and that at a time when they were still in full
use. The membersof the cult-communitywere unable to savethe sacredfur-
nishingsin the way in which other pagancult-furnishingswere savedfrom the
Christiansin the fourth century-byhiding them in the ground.The seemingly
suddenfall of the Dolichenus cultis thus possiblyto be connectedwith political
conflicts during the period of military anarchy.The so-calledMithraeumIII at
Poetovio,195luxuriously furnished by the Dacian legions under Gallienus, or
the Sol-cult busily propagatedby Aurelian, to name only two examples,are
significant for the entanglementof politics with religion.
Christianity probablydid not makeits appearance in the middle Danubearea
beforethe time of Gallienus. The first fairly secureindicationsof the existence
of communitiesof Christiansoriginatefrom this time, but this of coursedoesnot
excludethe possiblepresenceof individualsearlier.During thepersecutionunder
Diocletian,a numberof martyrsduring their trial or whenspeakingto believers
referredto their having beenbrought up as Christians;but the possibility can-
not be excludedthat they were born as Christiansnot in one of the Danube
provinces but somewherein the east, particularly since the Christiansearliest
attestedon the Danubemainly camefrom the Greeklinguistic area.Before the
Tetrarchy,Christianity seemsto have been strongestin and around Sirmium,
and this was presumablyconnectedwith the increasingnumber of foreigners
present there from the middle of the third century. The leading Danubian
soldiersand politicians were all paganswith an often markedhostility towards
Christians,which was evidently rootedin their old Romanattitude.In the third
century Christianity was not yet a political factor in the Danubearea.
Latin and Greek elementsin provincial culture

In the civilization of the Severanperiod, as alreadymentioned,the local traits
producedby the Romanizationof the secondcentury yielded to a colourless,
undifferentiated,empire-wide culture. That this culture was Latin-Roman in
almost all its aspectsand that the oriental elementswere representedin it only
as far as thesehad penetratedRoman civilization in the westernprovincesin
general,is due to the fact that Upper Moesiaas well as Pannoniabelongedto the
so-calledLatin provincesof the empire. Latin and Greek provincesto all ap-
pearancewere strictly separated,andthis separationprobablydid not correspond
to the linguistic situation in real life-first, becausethe local languages(Celtic,
Illyrian and Thracian),as we shall seebelow, were not given up by the original
259
population, and, second, becauseGreek was probably much more widely
distributedas a colloquial languagethan appearson the inscriptions,which are
with few exceptionsin Latin.I96 So we may be justified in supposingthat the
separationcame from above, and was official. The demarcationline between
Latin and Greek(often calledthe Jirecekline after the BalkanologistK. Jirecek
who first mappedit) ran along the southernand easternboundariesof Upper
Moesia. Macedoniaand Thracewere Greek, Upper Moesia-andeven more of
course Pannonia-Latinprovinces. The fact that this line coincided with the
frontier of the provincesclearly proves thatthe official languageof eachpro-
vince had been decided some time during the early imperial period. In this
respectit is noteworthy,too, that the borderstationwherethe most important
route in the Balkans,the great diagonalByzantium-Singidunumroad, crossed
the Jirecekline was calledLatina.I97 Thosecomingfrom 'Greek'Thraceentered
'Latin' Moesia at this point.
Upper Moesia and Pannoniawere Latin provinces in the sensethat the
languageof public life and of administration-bothat provincial and at munici-
pal level-was Latin. Hence the languageused for the inscriptions, whether
official dedicationsby imperial officials and municipal dignitaries,or appearing
on private votive andfunerarymonuments.This of courseby no meansimplies
that only Latin was spokenin towns and fort-settlements.In the larger settle-
ments a presumablyconsiderablepart of the- population spoke Greek for cer-
tain; this is clear above all from Greek graffiti on tiles and on the walls. The
non-Latin elements,however, erectedLatin inscriptions in those places,too,
where the public apparentlyunderstoodanotherlanguagebetter. The Syrian
families at Intercisaand Brigetio set up neither Syrian nor Greek inscriptions,
and at Scupi, where the populationof the town was probably better versedin
Greek than Latin, the languageof the grave-inscriptionswas Latin too. In
Scupi the Greek languageturns up only in unintendedslips, for instanceon
some inscriptions where the Latin text was renderedin Greek letters,I98 evi-
dently becauseit was dictated to somebodycapableof writing only in Greek.
On a numberof inscriptionswherethe essentialpart-thenameand age of the
deceased-was in Latin, the subjectiveadditions (verses,expressionsof grief,
etc.) were addedin Greek for the reasonthat they could not be producedin
Latin by the membersof the family who had hardly any commandof that
language.I 99
It is very likely that Greekwas particularly to be found in the largesttowns,
the very places in which Latin was most used as the colloquial language.A
broad class of the populationspeakingonly Latin or mainly Latin can be sup-
posedin towns only wherethe populationconsistedof a homogeneousmassof
260
the descendantsof Italian immigrants, or else of a heterogeneousmixture of
immigrantswho could makethemselvesmutually understoodonly by meansof
a neutrallanguage.And for this reasonLatin never really madeprogresswhere
the great majority were natives; the reasonwas that the languageunderstood
by all was alreadyprovided. Thus it will not be possibleto supposetoo wide a
distribution of Latin in the country and in the less important towns, especially
those of late foundation; still less a section of the population with Latin as
its native language.The fact that we neverthelessdo not possessany inscrip-
tions in local languagessimply meansthat writing, like many other attributesof
higher civilization, was an inseparablepart of the externalsdependingon and
producedby Romanrule.
The fact that the representativesof the Illyrian soldiery who had their origin
in the native Pannonianpopulation understoodonly one of the two cultural
languagesof the empire,namely Latin, is unequivocallyattestedby a source.200
But the majority of Pannoniansoldiers did not even speak Latin. Grave-
inscriptionsset up by Pannoniansoldiersof the Romegarrisonare often so full
of mistakesthat they form eloquent testimoniesto the defectivenessof their
commandof the Latin language.20r All the more attentiondeservesto be paid
to a specialgroup of Pannonianmonumentsin stone,attestinga knowledgeof
themes from classical mythology. Mythological subjects were popular on
gravestonesin the secondcenturyin westernPannonia202 and at Viminacium203
too (PIs 6c, lIb); but in west and south Pannoniaand in Upper Moesia they
were not so widely distributed as in east and north Pannonia,where from the
middle of the secondcenturyonwardsthey formed a very specificandnumerous
group (PIs 32, 33).204 The reliefs were mostly parts of larger, sometimeshuge
grave monuments.The mode of representationand compositiongoes back to
well-known and widespreadthemes of Hellenistic-Romaniconography,and
the variety of themesis remarkablylarge: it extendsfrom the old-Romanlegends
of Aeneasand of the foundation of Rome to little-known stories from Greek
mythology. In north-easternPannoniait had alreadybeena fashion before the
Marcomannicwars to furnish grave monumentswith stone slabs ornamented
with reliefs; theseearlier reliefs, however,took their themesfrom the everyday
life of the deceased,and a few very lively and attractivepopularrepresentations
resultedfrom it;205 the most delightful is a representationof a womanin native
dress,holding a pig's headon a woodenplate (PI. 1 3a).206It seemsvery charac-
teristic of the culture of the Severanperiodthat this popularart was given up in
favour of a classicalone. In part the classicalthemesare closely connectedwith
the cult of the deadand with conceptionsof the other world, in part they can
be attributedto the self-assurance,mentionedbefore, of the inhabitantsof the
261
The secondage oj prosperity: rise and co/lapse
Danubecountries,which expresseditself in old-Romancults andpractices.The
apparentlyvery popular sceneof Aeneasfleeing from Troy (PI. pb), the scene
of Rhea Sylvia with Mars or of the she-wolfwith the twins at any rate fit well
into the frameworkof this theory. Perhapswe may mentionherefurther Trojan
scenessuchas Priam beforeAchilles (PI. 3p), the draggingof Hector (PI. 3 3b),
or the story ofIphigenia;but the majorityof thesescenescanhardlybeexplained
in terms of old-Romanself-assurance. Of course,the questionhas also to be left
open whether all these scenes,which are sometimeseven representedin an
abbreviatedand simplified manner, were understoodand recognizedby the
customersandpublic at all. It is quite possiblethat the sceneswere simply taken
from the pattern-booksof the stonemason'sworkshopswithout any scholarly
understanding.Perhapswe should only stressthe supersessionof scenesfrom
everydaylife by thosefrom classicalart; this new choiceof themespresumably
indicates consciousadaptationto classical culture, an adaptationpresumably
basedon political and sentimentalmotivesbut by no meansimplying the spread
of deeperknowledge.In this respect,too, societyin the frontier region wished
to inflate its image; it is certainly not coincidental that the classical themes
were most widespreadpreciselyin the border area,wherethe populationdrew
most of its gain from prosperitybasedupon military policy.
The existenceof schoolscapableof providing morethan reading,writing and
arithmetic can be supposedonly in the capital towns; someevidencesuggests
that the sons of the small-town aristocracyhad to go to school in Carnuntum
or Aquincum (PI. 29b).207 Instruction in law can also perhapsbe assumed,to
provide the elementaryknowledge necessaryfor administrative staff. Since,
however,the staff of the provincial administrationwere soldiers,it is not abso-
lutely necessaryto supposeelementarylegal training amongcivilians. There is
much to suggestthat knowledgeof the law and a certaininterestin the stand-
ards and terms of Romanlaw were derived from the military.208
A few quotationsfrom Vergil among scribbleson tiles 209 point to texts of
classicalauthorspossibly being read in the higher forms in schools,but wider
deductionscan be drawn neitherfrom theseVergilian quotationsnor from the
verses on graves which contain manifold Vergilian reminiscences. 2ID These
poemswere mostly only adaptedon the spot and the Vergilian flourishes curi-
ously derive almost entirely from books I and VI of the Aeneid;the authorsof
theseverses,not necessarilynativesof the Danubianarea,thus had a schooling .
not much superiorto that of their clients in the Danubeprovinces.2II
Illiteracy canbe demonstrated by meansof incorrect,often evenunintelligible,
grave-inscriptions.Since those who set up the grave monumentsundoubtedly
belongedto the upper classof provincial society, educationand social position
262
The secolldage oj prosperiry: rise alld collapse
must be treatedas not inseparable.It is known that it was not necessaryfor
soldiersto know how to read and write. The knowledgeof writing was indis-
pensableonly for thosewho wereemployedin the administrationandin business
life, and that is why the relatively best-educated
peoplemust be soughtamong
the municipal scribae,the non-freeemployeesof the financial administrationand
the administrativegradesamongthe troops. A freedmanof the Furii at Ulpia-
num, Furius Alcimus, put a legal question to the most famous jurist of his
time,2I2 though the decurions of Ulpianum themselveswere not necessarily
literatepeople.This point must be stressedin order to preventthe large number
of Severaninscriptionsbeing consideredas proof of the spreadof Latin civiliza-
tion-writing and language.2I3 An essential,if not the most significant charac-
teristic of the Romanizationof the Danube areas was that it took on forms
which did not correspondto its real content,and that in all its manifestations
it exaggeratedits intensity. This, however, was only possible in the special
circumstancesof a politically basedprosperity. Certainly the Romanizationof
the secondcentury, which had createdin many areasof our provincesa special
culture of local colouring,was much more honestand startedon a development
which later on might have contributedsomethingof value to the civilization
of the empire. But when the soldiery suddenlyrose to political importancethey
wishedto stresstheir Romanidentity at the expenseof thesespecialtraits, and
thereforecreatedthe colourlessDanubianRomanizationwhich, althoughdraw-
ing much from the culture of the empire, lacked a deeperunderstandingof it.
This very superficial and in many respectsonly make-believeRomanization
could not hold its own againstthe storms which were soon to break out, and
collapsedmuch more quickly than the Gallo-Roman,African and other cultures
in the empire which more successfullypreservedtheir local traits.
Third-centu~
collapse
To speakof a collapsein the true senseof the word is justified, not only be-
causethe barbarianinvasionsand wars in the secondhalf of the third century
were truly catastrophicfor the Danubeprovinces,but becausethere must have
been further, deeper causesresponsiblefor Pannoniarecovering so slowly
after the crises.To point to the generalcrisis of the empireis too easya way out.
But, in the last resort this generalcrisis does seemto have been the causeof
many of the barbarianinvasions.As we havepreviously mentioned, peaceon the
Danubecost the empire a presumablyconsiderablesum of money. This sum
was regularly paid by the Severi, becausethe availability of the Danubianarmy
for wars in other parts of the empirewas a considerationof the first importance
2. 63
with them, and the loyalty of the troops themselvesalso dependedto a con-
siderable extent on governmentwillingness to raise the sums necessaryfor
securingpeacealong the Danubefrontier. Although it goesbeyondthe author-
ity of our sources,it can be consideredprobablethat the elevationof pretenders
of their own by the Danubeprovincescoincidedwith inability to pay the annual
sums due to the barbarians,and this could occasionallylead to barbarianin-
vasions.The war underRegalianusagainstthe Sarmatianscould be explainedin
this way; and, vice versa,if the central government with the financial resources
of the empire at its disposalfor somereasonor other did not make the annual
moneyavailable,the Danubiantroops will havetried to put their own candidate
in control of the empire. The appearanceof the Goths on the lower Danube
mustnot be forgottenhere,for this was an entirely new factor in Romanforeign
policy which further aggravatedthe potentially dangeroussituation.
The war under Gallienus (259 or 260) in Pannoniaand the war againstthe
Gothsin 268 in UpperMoesiaprobablyresultedin devastationsat leastas great
as those which occurred during the Marcomannicwar under Marcus: at any
rate the very large number of coin-hoardsfrom all parts of Pannoniasuggestsa
generalcatastrophe.But though the situation in Pannoniareturnedto normal
again fairly soon after Marcus, nothing can be said of an early reconstruction
or consolidationin the last third of the third centuryin Pannoniaand in Upper
Moesia. The suddendecreasein the numberof inscriptionswell illustratesthis,
andit is also striking that in the whole of the archaeologicalmaterialwe haveno
types or groups of finds which can be dated indisputably to this period. This
presumablyindicates not merely an absenceof much building activity and
reasonableluxury during this time, but decline of industrial production to a
minimum. Nor can it be said that an import-tradeworth mentioningexisted.
We haveinscriptionsfrom the time of Gallienusto that of the Tetrarchy,but
these are mostly either votive or building-inscriptions set up by personsof
high rank (such as legionary prefects and governors)or else milestones.The
building-inscriptions attest a certain amount of reconstructionat military in-
stallations.But inscriptions set up for private purposesand grave-inscriptions
are practically missing altogether.The few exceptionsare markedout by strik-
ingly primitive or simple execution, which when comparedwith the good
standardsof stonemasonsin the first half of the third century attest a sudden
decline of the workshopsand of the artistic ability of the stonemasonsthem-
selves. Demandmust have lessenedconsiderably.Sarcophagiand fairly large
grave monumentswere not being commissioned;the only attestedtype of
grave monument is the gravestone( stele) which was in most casescrudely
executed.214
264
The generaldecline had far-reachingconsequences for the further develop-
ment of the provinces.The damagesufferedby both peopleand land was not,
it seems,put right during the period from Gallienus to the Tetrarchy.21 5 The
reasonappearsto be twofold. First, individual emperorswere not able to con-
centratethe forces of the empire in the Danubeprovinces-andif they could,
the problemof Daciahadfirst to be solved.Second,it was easierto disregardthe
particularisticview-point of the Illyriciani-and in this the 'Illyrian' emperors
themselvesset an example,althoughthey did attemptto achievesomethingin
Pannonia. It is presumablysignificant that Probus, who had been born in
Sirmium, compelledhis own soldiers to take part in the economicreconstruc-
tion of Pannoniaand Upper Moesia.216 This was a new policy, entirely different
from the economic policy pursuedby the Severi. It causedthe death of the
Emperor Probus, but the idea of reviving the economy of the Danubian
provincesunder their own efforts was a step in the direction of the methods
which were to be usedby Galerius.Decline and the end of the economicboom
were closely connected,in that the central governmentin which the Illyriciani
themselveshad a part abandonedthe parochial outlook of the Danubearmy.
In Pannoniaand to a certainextent in Upper Moesia too, this new policy spelt
bitter decadesof misery and insecurity.It is, of course,a different question,and
one which the historian cannotanswer,whetherit was becauseof the general
economiccrisis or by a gradually achievedpolitical insight that the Illyriciani
cameto, or were compelledto, lay asidetheir long-establishedegoism.
265
Chapter 8
The Danube frontier in the late Roman period
Aurelian fell victim to a court conspiracywhose effect was to bring the aged
Tacitus to the throne. To him the Illyriciani stayedfaithful while he lived,! but
on his deathsoonafterwardsthe troopsonceagainproclaimedan emperorfrom
the Danubeprovinces.Probus hadbeenborn at Sirmium and belongedto the
immediatecircle of Aurelian, and he continuedthe political programmeof his
predecessor,a programmewhich now had nothing in commonwith the paro-
chial views of the older generationof Illyriciani. The political outlook of Probus
showedthe sametendencytowards universalismas that of Aurelian had done,
and in his policies he could presumablyreckon on the co-operationand under-
standing of the Illyrican junta; the political thinking of this group, as had
becomeincreasingly obvious from the last years of Gallienus, had grown to
include a concept embracingthe whole of the Roman empire. The hostility
towards the Christians shown by Aurelian fits into the framework of this
concept, as does his intention to end the economic passivityof the Danube
provinces,so that imperial policies should not have to dependon the parochial
outlook which controlled the demandsof the Danubearmy. The mutilation of
our sourcesforbids descriptionof this economicpolicy which we can traceonly
in a few events, as for instancein the accountsof drainageactivity and the
planting of vines around Sirmium and on the Upper Moesian Danubeunder
Probus.2 Thesewere presumablypart of a larger economicplan, sincethey were
demonstrablycontinuedby Galerius on a large scale. A curious votive stone3
of about the sameperiod from Pannoniarecordsextensiveprivate planting of
vines; the fact that they appearon a commemorativestoneshowsthat activities
of this kind were consideredat the time as politically important.
266
The Danubefrontier in the late Romanperiod
The men of the Danubearmy probably could not rise to this new policy. It
was not only a personaltragedyfor Probusthat he was murderedby his soldiers
on accountof it: the soldiers on the Danubegradually ceasedto sharein the
control of affairs; after Probus'murder the initiative for choosinga successor
did not come from the Sirmian troops but from the army of Raetia, and an
emperor was elected whom even the Historia Augusta could hardly call an
Illyricianus. When Numerian and Carinus, the sons of this short-lived new
emperor,had been killed by membersof the Illyrican junta, the new regime
of the Jovii and Herculii no longer relied on the Danubearmy. The decisive
victory over Carinus took place near Margum in Upper Moesia4 in 285, and
won for the Dalmatian,Diocletian, undisputedrule over the empire. Although
the co-rulers,whom he sooncalled in, were all menfrom the Danubeprovinces
-Maximianus from Sirmium, Galerius from Romulianumin the new Dacia,
Constantiusfrom Naissus-asif all Danubianprovinceswere to be represented
in the Tetrarchy-thepolicies which they were able to bring to fruition cor-
respondedwith the imperial outlook of the new generationof Illyriciani, and
not with the old ideas of the Danubianarmy, which could be deprived of its
political influence by meansof the new economicpolicies. However, Sirmium
managedto keepits leadingrole andexperiencedits greatestperiodof prosperity
as one of the principal imperial capitals. It is therefore not surprising that
natives of Sirmium and its surroundings continued to rise to the highest
administrativeposts of the empire and that the dynasty of Constantinewas
finally replacedby them, when Jovianus from Singidunum and Valentinian
from Cibalaewere successivelyproclaimedemperor.
The evacuationof Dacia did relieve the situation beyond the frontier to a
certainextent,but the resultswere tangibleonly on the lower part of the Danube
limes. The Goths who had taken possessionof Dacia after its evacuationno
longer exerted pressureon the limes of Moesia and the new Dacia. But the
diversionof their forces in the direction of the old Dacia createdan increasingly
serioussituation on the limes of Pannonia.We are very badly informed on the
wars that occurredunder Probus,Carus, Carinus and Numerian; the accounts
in our sourcescan best be interpretedin the context of a processleadingup to
the numerousSarmatianwars of Diocletian.
The Sarmatianscameunder pressure,as alreadymentioned(pp. 209 ff.), not
only from the north, after the Vandals' migration, but also from the east and
north-east,from the Gepidaeand Goths. This may have beenthe main reason
for their renewedattacks on the limes of Pannonia.The first attack, of which
we are informed only indirectly,S happenedunder Probusc. 278; it is possible
that the Vandals also took part in it. A few years later Carus gaineda victory
267
over the Sarmatians, 6 and about the sametime a victory over the Quadi was
celebrated. Since this victory can be dated to the year 284, and as Diocletian
7
took the title Germanicuson one more occasionthan his colleagueMaximian,

it is likely that this war againstthe Quadi continuedperhapsinto 285 and was
broughtto an end by Diocletian acting alone.8 The DanubianSuebiafterwards
causedfew problems,9the Sarmatiansmany more: the latter were to give the
Tetrarchyas well as the Constantiniandynastyplenty to do.
Unfortunatelywe have only a very rough idea of what took placein the Sar-
matianlowlands after the evacuationof Dacia. At first the Sarmatianswere able
to offer resistanceto pressureby the Germanic tribes, but their attacks on
Pannoniaprove their renewedattemptsto obtain armedassistancefrom Rome
or the receptio. Leaving aside doubtful archaeologicalfinds, good evidencefor
the receptio of a few smaller groupsof Sarmatians10 is afforded chiefly by place-
names,such as Mutatio Sarmatarumin the Morava valley,II and by the mention
of a few Sarmatianunits in the Notitia;12 but of these,some,at least,were pro-
bably admittedinto the empirelater.13 Becauseof the Sarmatianwars Diocletian
had to spendmost of the period 289-94 at Sirmium,i4 even after he appointed
Galeriusruler of Illy ric urn (on I April 293).15Only the outline of the chronology
of the Sarmatianwars is at all certain;16 accordingto it, victories over the Sar-
matiansoccurredin the winter of 289-90,in 292 and in 294, when an important
victory may havebeenwon; thenin 299 whenthe Marcomannihadto be fought
too; and three furtherSarmatianwars of Galeriusare attestedin the years 299-
305. Unfortunately we do not know whetherthe new assumptionsof the title
Sarmaticuswere the results of actual military successesor whether smaller
skirmisheson the bank of the Danubeand merediplomatic stepswere sufficient
to justify them. At any rate the victory of 294 received special celebration,17
and a few other eventsin the sameyear suggestimportant measuresand suc-
cesseson the part of the Romans.In the autumnof 293 Diocletianinspectedthe
limes of PannoniaInferior; his presenceat Lugio, one of the most important
stationson the Sarmatianfront, is attestedby two edicts issuedon 5 Novem-
ber.18 Since Galerius had then been Caesarand the real ruler of Illyricum for
six months, the presenceof the Augustus-hissenior in rank-in Galerius'
dominionsmust be connectedwith particularlyurgentandimportantmeasures;
Diocletian visited Pannoniaon only one further occasion,for the conference
in 308, when he was constitutionallyno longer emperor,I9The Sarmatianwar
of 294 was probably conductedby Diodetian in person: he remainedat Sir-
mium from September293 till August 294 almost without interruption. The
preparationsmade in the autumn of 293 were evidently intendedfor the war
plannedfor the following year, but the measuresbeing taken were quite novel
268
and so demandedthe presenceof the seniorAugustus.Thesenew measuresare
recordedfor us in an entry for the year 294 in the ConsulariaConstantino-
politana,which reads'This year forts werebuilt in SarmatiaoppositeAquincum
and Bononia' (his consulibuscastrafacta in Sarmatia contra Acinco et Bononia).20A
seriesof coins with the new reversetype showing a fort-gate, which was to be
repeatedoften in the future, may also be relatedto this achievement. 21
The two
forts are also mentionedin the Notitia Dignitatum as garrisonsin barbarico 22
which is all the more curious since counter-fortificationson the left bank of the
Danubeand on the right bank of the Rhine are never otherwisecalled forts on
barbarianground in this source. In the case of these two forts, therefore, in
barbarico must meanthat they were built not on the limes but deepin the country
of the Sarmatians.
The fort-sites havenot so far beenfound; but a few indications,amongthem
a large votive inscription of Diocletian and his colleaguefrom the Lowlands
reading ob devictos virtute sua Sarmatas23 (according to R. Eggers'srestoration
of the text), strongly suggest the existence of Diocletianic fortifications.
Furthermore,though researchon the Roman frontier has made considerable
advancesin Hungary it has not so far succeededin proving a Diocletianic
period of construction on the Danube limes, although large numbers of late
Roman frontier buildings, as we shall see, have been discoveredand investi-
gated.Finally we find on the Sarmatianbank of the Danube,the ripa Sarmatica,
a type of fortification which itself suggestseither that Roman troops were
permanentlystationedin Sarmatianterritory or that constanttransportof sup-
plies thither was necessary.
This type of fortification 24 has so far been attestedat about sevencrossing-
places,on the right as well as on the left bank. It is a rectanglesurroundedon
three sides by walls and towers and on the fourth open towards the Danube
(Fig. 42); the walls run down into the water. This type of plan can be inter-
preted as a fortified landing-place, and since a special tile-stamp, which is
attestedon the Sarmatianlimes only in thesefortifications, suggeststhat they
were built at one time, we must either connectthem with a plannedoffensive
or preferablywith a phaseof foreign policy which necessitatedregular inter-
course betweenPannoniaand the Sarmatians.The obvious conclusion is to
connecttheselanding-placeswith the constructionof fortifications that occurred
in 294. If this rather bold reconstructionof eventsis correct, then the absence
of a Diocletianic building period on the Pannonianlimes can be explained.
Diocletian and Galerius presumably recognized that the Sarmatiansunder
pressurefrom the north and easthad no choice but to force a receptio or armed
aid by continual attacks.Under such circumstances,reconstructionand further
269
Danube
0, soI AQVINCYM
Rakospatak
metres
AQ V IN
Rakospatak
C Y M
A Q V IN C A Q V IN C Y M A Q V I
Hora.ny
YM
NC
M Y
CONTRA FLORENTIAM
Figure 42 Fortified landing-places

fortification of the Danubelimes could only have beena palliative, and for that
reasonconstructionon the limes was abandonedand an entirely new measure
in the field of foreign and military policy was introduced:the partial occupation
of Sarmatiancountry by Romantroops.The aim was not conquestor formation
of a province but simply to ensuresecurity and peaceon the Danube.It could
also be interpretedas armedaid for the barbarians.Eventson the ripa Sarmatica
in the first half of the fourth century thus lead us to the conclusionthat the
Sarmatianswere bound to Rome by special treatiesand that the disturbances
which broke out from time to time were causednot by fundamentalhostility
on the part of the SarmatianstowardsRomebut by pressureon them from their
neighbours.
It would be tempting to date the so-calledDevil's Dyke (Csorszarok)in the
Sarmatianlowland to this period (Fig. 43). This earthwork, which has only
A Q V IN C Y M
A Q V IN C Y M AQ
VI
AQV
NC
YM
INC
AQV
AQYM
VI
IA
NQ
NCY
CV
YIM
YM
NCY
M
C
V IN
M
AQ
O
.
L-. .
k k km
Figure 43 The 'Devil's Dyke'

271
recently been mappedand investigated,25enclosedthe very area inhabited by
the Sarmatiansin the third and early fourth century. We alreadypossesscertain
clues for dating it, notably a terminuspost quemin the third century: the earth-
work in someplacescuts through tracesof settlementsof the secondand third
centuriesandthereforecannotbe datedbeforethe endof the third centuryat the
earliest. Other finds prove that it must have been built before the ninth cen-
tury.26 As a fortification it had hardly any value at all, but as a visible boundary
line againstthe peopleson the fringe of the Lowlands, the Vandals, Gepidae
and Goths, it could playa certain part, provided it could be somehowguaran-
teed by a third power. As long as its exact date remainsunknownit is wiser to
leavethe questionundecided.A slightly later datewas recentlyproposedby the
archaeologistinvestigatingthe earthwork.27 The dyke must have been in use
for a long time, sinceits coursewas altereda numberof timesand somesections
showedevidenceof reconstruction. 28
That the new governmentof Galeriusat Sirmium consideredthe war of 294
and the measurestaken after it as a solution of the Sarmatianproblem is clear
from the fact that he soonafterwardssetout againstthe Carpi and otherenemies
on the lower Danubelimes and settledpart of the defeatedpeoplein Pannonia.
Moreover,thesesettlementswere connected with large-scaledrainage,and soon
after they occurredthe Diocletianic administrativereforms cameinto effect. All
this suggeststhat after 294 Galerius consideredthe situation in Pannonia
sufficiently consolidatedfor far-reaching reforms.Only 299 saw further war,
and this probably took place on the northern border of the Sarmatiansand in
the territory of the Marcomanni.29 OtherwisePannonia,Upper Moesia and the
new Dacia were able once more to enjoy an assuredpeace.
In the yearsafter 294 Galeriusmaintainedthe economicpolicies on the middle
Danubewhich had been started by Probus. Drainage was carried out in the
interior of Pannoniaby canalizing the natural outlet from Lake Pelso to the
Danube(i.e. by building what is today the Si6 canal) and this, like the drainage
around Sirmium undertakenby Probus, was designedto increasethe area of
cultivatableland.3D At the sametime the Carpi were settledin Pannoniaafter
their defeatin 295 on the lower Danube.Possiblythesebarbariangroupswere
given land in the very areasthat had been drained, as had happenedwith the
Cotini, who hadbeensettledbyMarcusaroundMursaandCibalae,in the marshy
region of the Hiulca paluso Possiblythe improvementsthemselvesmay havebeen
carriedout by the newly settledbarbarians.A gift of land to the Carpi in eastern
Pannonianorth of the Drave can be establishedfrom a passagein Ammianus
Marcellinus.31 The Si6 canal flows through this area, and it was this part of
Pannoniawhich was createda new province under Galerius.
272
The late Romandivision of provincesand separationof civilian and military
administrationwas applied only gradually in our area. The first step had been
taken by Aurelian when he establishedthe new Dacia. Of the towns of Upper
Moesia, probably only Ratiaria and Aurelianum at first belongedto this pro-
vince. After 294 underDiocletian, or more accuratelyunder Galerius,the great
reform was set in motion. As we have stated, thenorth-easternpart of Pan-
nonia, that is to say the northern half of PannoniaInferior, was constituteda
new province. The date at which this occurredcan be establishedfrom the fact
that the new province was called after Valeria, Diocletian's daughterand wife
of Galerius,and from the associationof naming and drainage.32 Thus the pro-
vince of Valeria cameinto existenceafter the victory over the Carpi, i.e. in 295
at the earliest.33
A few uncertainindications34 suggestthat the remaining late Roman pro-
vinces of Pannoniaand Upper Moesiawere organizednot then but later, and it
is not impossiblethat the administrativereforms of the period from Diocletian
to Constantinethe Greatwere modified to somedegree.In the absenceof direct
information, and also of thoroughresearchinto thesequestions,it will be better
to describehere the final situation as it had come to exist roughly under Con-
stantine.
In Pannonia,four provinces came into being as a result of the reforms
(Fig. 44): Valeria was formed from the northernhalf of PannoniaInferior, and
PannoniaSecundafrom the southernhalf of the sameformer province, while
Savia was constitutedfrom the southern,and PannoniaPrima from the north-
ern part ofPannoniaSuperior.The borderbetweenUpper andLower Pannonia
from now onwardsis that betweenPrima and Valeria, and Savia and Secunda
respectively;the frontier betweenSaviaand Primafollowed the Drave, but that
betweenValeria and Secundaran north of the Drave, sincea pieceof land north
of Mursa-presumably the territorium of this town-belongedto Secunda. 35 The
seatsof the civilian administrationwere the towns Siscia in Savia, Savariain

Pannonia Prima, Sopianae in Valeria and Sirmium in Pannonia Secunda.
Sirmium also becamethe centreof a unit of administration,being variously the
capital of a dioceseand of a prefecture.The governorsof the Pannonianpro-
vinces were not equal in rank, but their status was altered during the fourth
century,and all that is certainis that, when Romanrule cameto an end, Prima
was undera praeses,Saviaundera correctorand Secundaundera consularis.36
In the courseof the administrativereforms a town of Pannoniawas cededto
Noricum: under Constantine,Poetoviono longer belongedto Savia.37
In the areaof Upper Moesia the following provinceswere established(Fig.
45): MoesiaPrima and Dardanialay wholly within the former Upper province;
273
PAP
NAN
.SAVARIA
NONN
PANNONIA
OIN
PRIMA
AIA
PA
NN ~OPIANAE
ON
PIA
NO N
PRIMA NIA PA
PANN
NNON
ONIA
IA ~OPIANAE
o 50 100 k
.'....__f~====~1 m
Figure 44 The late Romanprovinces:Pannonia
DaciaRipensisandDaciaMediterranea,however,extendedinto partsof Moesia

Inferior (the remaining area of which was named Moesia Secunda)and into
Thrace.It can,however,be shownwith a large degreeof probability that Aure-
lian's new province of Dacia and Dacia Ripensishad the sameextent. At what
datea secondprovincenamedDacia was establishedis not certain,but possibly
not before Constantine,when Dacia Mediterraneawas createdout of parts of
Dardaniaand Thrace.For Naissus,which certainly belongedto Dacia Mediter-
raneafrom the time of Constantineonwards,is still mentionedas a town of
Dardaniaonly a few yearsearlier.38 This, incidentally, was the reasonwhy the
family of Constantinecould be traced back to the Trojan Dardanos.Dardania
and Moesia Prima were constitutedby dividing the remainingareaof Moesia
Superior,probablyunderDiocletian.Later, the territory of Naissuswasattached
274
/MOESIA
/M
/MOESIA
O
/MOESIA
E S/ M
IAOE
/M /M
OE
SI
OE
SI
A
/M A SI
A
OE
S I/ M
AO
ES
I A/ M
OE SERDICA
SI
A
/MOESIA
/MOESIA
CVP1
o
, so
(
100
I
km
Figure 4J The late Romanprovinces:Upper Moesia
275
The Danubefrontier in the late Roma/lperiod
to the new-foundedprovince of Dacia Mediterranea,and this arrangementwas
maintained until the time of Justinian. Dardania included the territories of
Scupi, Ulpianum and municipium Dardanorum;the parts of the two Dacias
that formerly belongedto Upper Moesia were the territories of Ratiaria and
Aurelianum and of Naissusrespectively;the rest was given to Moesia Prima.
The seatsof the provincial administrationwere at Viminacium, Ratiaria, Serdica
and Scupi. In the Notitia Dignitatum the governorsof the provinces,with the
exceptionof the consularisof Dacia Mediterranea,are all praesides.3 9
The ten yearsuntil the abdicationof Diocletian in 305 probablypassedwith-

out any notablehostilities on the Danubefrontier. The written sourceson the
reigns of the last paganemperorsare very incompleteand distorted-presum-
ably this is no coincidence-butChristiantradition would not have suppressed
wars of any importance conducted by Galerius. The economicreconstruction
under Galerius,which is briefly alludedto by later historians,also suggeststhat
the situation at that time was probably peacefuland stable.The victorious title
Sarmaticus,which was frequently adoptedby Galeriusand his successors,must
be understoodin the sameway as the titles sempervictor, triumphator, etc. The
adoption of the title Sarmaticuscould be the result of diplomatic moves or
simply of the emperor'ssuccessin maintaining peaceon the ripa Sarmataca;
in this sensehe would have beenthe conquerorand masterof the Sarmatians.
Moreover,Galeriushadto fight severalbattlesin the eastand, after296, was not
personallypresentmost of the time on the Danubianfront. 40 A war againstthe
Sarmatiansis attestedduring the secondhalf of 305, in which the young Con-
stantine' accompanyingGalerius as Caesar,distinguishedhimself.4I The scale
of this war, whoseonly recordedepisodetook placein a marshyareaand there-
fore presumablynot far from the Danube,cannotbe established.
The long-drawn-outcrisis of the Tetrarchyfrom the abdicationof Diocletian
to the assumptionof solerule by Constantineaffectedour provinceonly towards
its close. Galerius, the only consistent representativeof the Illyrican junta
among the Augusti and Caesars,was undisputedruler on the middle Danube,
and when the emperorsassembledat Carnuntumin November30842 Illyricum
onceagaincameunderthe rule of an Illyricianus. However,the conferenceof the
emperorsat Carnuntumwas indicative of the changedbalanceof power in the
empire inasmuchas the choice as venue for the negotiationsfell not on Sir-
mium, the stronghold of the Illyriciani, but on what was by now a rather
unimportanttown in Pannoniawhich was not even the seatof a praeses.When
choosingthe place for the meeting of the religiosissimi Augusti et Caesares-as
they called themselveson the altar to Mithras (PI. 34b) from Carnuntum43_
Sirmium was evidently deliberatelyavoided.
276
The Danubefrontier itt the late Romattperiod
In November308 the PannonianandUpperMoesianprovinceswereassigned
to the rule of Licinius, a fellow-countryman of Galerius.44 At first Licinius
probably residedfor most of the time at Sirmium, but his court had to leave
there atthe end of 314. ConstantinereachedPannoniawith his army in Septem-
ber of that year, defeatedthe more powerful army of Licinius on 8 Octobernear
CibalaeandpursuedLicinius down the Balkan peninsulaas far as Thrace.45 As a
result Constantinebecamethe virtual master of our provinces, though the
leading circles of Illyricum were presumably opposed to him. At Cibalae,
Danubiansoldiers had been defeatedby west Europeantroops, and a curious
group of hoards from Pannoniaand Moesia suggeststhat the Licinian party
continuedstrongand that Constantinefound himselfconfrontedwith consider-
able internal political difficulties in the region of the middle Danube. Silver
disheswith the punchedinscription Licitti Augustesempervittcas (PI. 35b), which
wereproducedon the occasionof the tenth anniversaryof Licinius' accessionin
318, apparentlyin the court workshop, and then were sent to personsof high
rank filled with gifts, havecometo light in severalplacesin Hungary,Jugoslavia
and Bulgaria.46 They prove two things: first, a busy propagandistactivity on
the part of Licinius in the dominionsof Constantine,and, second,a persecution,
sooneror later, of the followers of Licinius; for the disheshave beenfound in
fairly large numbersonly becausethey had to be concealedin the ground.
Between 315 and 324, that is until the elimination of Licinius, Constantine
stayed practically continually at Sirmium; the interruptions were short visits
to north Italy (315-16, 318) andin particularto the middle Balkanarea,wherehe
is attestedmostly in Serdicaand in his birthplace, Naissus(31 6- 1 7, 319, 320,
321, 322, etc.).47 Serdica gradually becamea rival to Sirmium. In Sirmium,
Constantinepresumably met much opposition from the Illyriciani, among
whom there were many supportersof Licinius. His long stay at Sirmium was
presumablyconnectedwith this, since there were no reasonsof foreign policy
for his staying: the Sarmatiansbecamehostile only from 322 onwards. The
allegedintention of Constantineto chooseSirmium as the capital of the empire
was possiblyonly a rumour spreadfor tactical reasons,with an eye to the Illyri-
ciani. After the removal of Licinius and especially after the foundation of
Constantinople,Constantine'svisits to Sirmium were limited to the occasions
of passingthrough on his way from the Balkansto the west.
In Mayor June 322 Sarmatiantroops under the leadership of a certain
RausimodusinvadedValeria.48 The chanceof tradition allows us to identify the
site of the battle: Rausimodusbesiegedthe Danubianfort of Camponanot far
from Aquincum and set the wooden superstructureof the stone fort on fire.
Constantinesoonappearedfrom Sirmium, attackedRausimodusnearCampona
277
and pursuedhim beyondthe Danube,wherehe defeatedthe Sarmatiansin their
home territory and Rausimodusmet his death. This was an opportunity for
Constantineto undertakean expedition through the whole of Sarmatia. He
enteredimperial territory again near Margum in Moesia and thus presumably
traversedthe whole of their land from north-westto south-east.The reasons
for this seeminglyconsideredundertakingare not quite clear. Constantinespent
hardly more than a month beyondthe frontier and soonwent to Savaria,where
he is attestedby the end of July.49 His whereaboutsin summerand autumnare
not recorded,but at the end of the year he was in Serdica.It is not impossible
that the journeys of the emperorin the summerwere connectedwith the pre-
ceding operationin Sarmatia.According to OptatianusPorfyrius, Constantine
brought back much booty, especiallycaptiveswho were distributed at Bono-
nia.50 It is quite possible that in 322 Sarmatiangroups were settled in the
Danube provinces and that a new treaty was made with their compatriots
beyondthe frontier.
This suggestionfinds its confirmation in the struggle which broke out ten
years later, for this allows an inference about events in the Danube-Tisza
plain between294 and 332. Sinceparts of Sarmatianterritory, as we have seen,
were occupiedby Roman troops, and since these garrisons,according to the
Notitia werepermanent,we cannotregardthe action of Rausimodusas a simple
plundering raid, but rather as an indication that peaceon the ripa Sarmatica
could no longer be securedin the mannersuccessfullyusedin 294. The attack
by Rausimodusmust have beencausedby increasingpressurefrom the Goths,
for the Romanvictory over the Goths in 332 within Sarmatianterritory 51 can
only mean that before this date the Goths had already invadedan area which
was officially recognizedas Sarmatia.The testimonyof a further sourceallows
us to infer that in 332 the Sarmatianswere allies of Rome,52 evidently as a result
of the treaty renewedin 322. Sincewe are further informed that after 332 a huge
massof Sarmatianswas settledin the Balkansand in Italy,53 we must interpret
theseeventsin the light of the old alternative: armed aid or receptio, the root
causebeing increasingpressureexerted on the Sarmatiansby the Goths and
other Germanictribes. When in 322 Constantinemarchedacrossthe country of
the Sarmatianshe might well have convinced himself that they were caught
betweentwo fires andthat their problemcould be solvedonly by effectivearmed
aid or by receiving some of their groups into the empire-i.e.by receptio. To
evacuateSarmatiaaltogetherand thus to sacrifice it to the Goths, Gepidaeor
Vandals was apparently a solution that could not be seriously considered.
Constantinethereforehad no choice but to transfer Sarmatiangroups into the
empireand at the sametime to provide effective armedassistancefor thosethat
278
remainedoutside. The Devil's Dyke which cuts off Sarmatiancountry towards
the north and the east will have beenrecognizedby the Romansas a frontier,
and very possibly it was under the formal supervisionof Roman garrisonsin
Sarmatia;this produceda commitmentwhich Roman policy could no longer
avoid in the old manner.
This is the reasonwhy in 332, when the Goths penetratedinto Sarmatia,the
Romansmadea decisiveand energeticinterventionon the Sarmatianside and
'defeatedthe Goths in the country of the Sarmatians'.54This success,aided by
large-scalefortifications on the lower Danube,could have forced the Goths to
keep quiet for sometime, if the Sarmatianshad remainedcontentwith Roman
help. Insteadthey took a desperatemeasurethat brought its own immediate
punishment,by arming their subjects,the Limigantes. After the victory over
the Goths an uprising of the armed Limigantes broke out, and part of the
Argaragantes,the Sarmatianruling class,had to flee into the empire.55 Another
part of the Argaraganteswere given refuge by their northern neighboursthe
Victohali.56 Thenfollowed two yearsof war with the insurgentSarmatians,and
Constantine'sown presencethere in 334.57 A huge coin-hoard at Campona
suggestsheavylossesamongtroops on the frontier. 58 In a descriptionof much
later eventsin Ammianuswe learn59 that the Limiganteswere organizedunder
Roman supervision in the border section of PannoniaSecundaand Moesia
Prima. They were divided into two groups,namedafter two Romanforts: the
Amicensesafter Acumincum (Slankamennearthe mouth of the Tisza) and the
Picensesafter Pincum(Veliko Gradiste); it seemsthat control over them was in
the handsof the commandersof theseforts.
The plain, which had becomesomewhatdepopulatedas a result of thesewars,
and particularly owing to the emigrationof the Argaragantes,becamea boneof
contention among Germanic tribes; the wars between individual tribes are
reported-withmany gapsand in a mutilatedform-only by Jordanes.Despite
the Romanvictory over the Goths,the latter, andperhapsthe Gepidaegradually
too, probably becamemastersof large stretchesof the plain. The Vandal king
Visumar, who likewise laid claim to the depopulatedcountry of the Sarmatians,
was defeatedby the Goths and fled to the Romanempire.60 In the easternhalf
of the plain, particularly in the area north of the river Maros and east of the
Tisza, Germanic influence becomes perceptible in the finds of the fourth
century,6I and this suggeststhat the Devil's Dyke gradually lost its role, at
least in that part facing the original Dacia now occupied by the Goths and
Taifali.
It is very likely that the transferof Sarmatiangroupsinto the empirewent far
wider than reception of the expelled Argaragantes,and was deliberately and
279
systematicallyundertaken byConstantine.For we have records of a mass of
peoplein unprecedented numbers,who were settledin Thrace, Scythia Minor,
Macedoniaand Italy, whilst more than twenty years later we hear of Argara-
ganteswho had remainedbeyondthe frontier.62 The war of 332-4 thus displays
certain similarities with the Marcomannicwars: the enemy was defeatedbut
receptio was also employed. Constantinehad presumablyrecognizedthat force
and conclusionof treatieswere not by themselvessufficient when the peoples
in the border lands outsidethe frontier cameunder pressurefrom beyond.
The combinedmethodof receptio and military interventionnow resultedin a
longer period of peaceon the middle Danube.From c. 335 to 356 no disturb-
ancesworth mentioningtook place on the ripa Sarmatica.Constantinepresum-
ably did not neglectto negotiatewith the Germanictribes who had invadedthe
Lowland; the treaty with the Goths on the lower Danube no doubt had its
parallel on the ripa Sarmatica.
All our evidenceso far supportsthe view that late Romanfortifications on the
ripa Sarmatica,new buildings and reconstructionsalike, were first begununder
Constantine.We have alreadypointed out the absenceof a Diocletianic period
of constructionin Pannonia,and we have connectedthis with a new frontier
policy inauguratedby Diocletian, with which the Devil's Dyke on the plain
may be associated.If this hypothesis-over-bold in the presentstateof research
on the earthwork-is correct, it follows that the very marked Constantinian
period of constructionon the Pannonianlimes means a return to traditional
frontier policies of the pre-Diocletianicperiod. The eventsof the years 332-4-
if their interpretationis correct-maythrow light on this return from a different
point of view.
The periodsin the history of the Upper Moesianlimesare less clear. The very
intensivenew work on the limes in the Djerdap has produceda numberof new
fort-plans, but a definitive identification of their periods must be reservedfor
future work (Fig. 46). The situation is renderedmore difficult by the further
fact that fortification of the Iron Gates was also undertakenunder Justinian.
Furthermore,this particular section of the limes was least accessiblefrom the
left bank, and so the absenceof certaintypes of tower by no meansprovestheir
absenceon other parts of the limes in Moesia Prima and Dacia Ripensis.More-
over, a kind of Devil's Dyke is also known in the country opposite Dacia
Ripensis: this is the Brazda lui Novac, which runs eastwardsfrom the Iron
Gatesand is designedagainstan enemyfrom the north.63 It could equally well
be attributedto a time when placeson the left bank of the Danubewere not yet
given up after the evacuationof the old Dacia and thereforehadto be protected,
as to the activities of Constantinewho had bridgesconstructedand roads built
2.80
Danube
6
TALIATA
3
Bosman
1 NOVAE Cezava
5 Ravna
6 TALIATA
50
o.I________ '~====~I
700
metres
Do~i
6 TALIATA
Mi lanavac (,}
6
.0
6 6
•
1 6 6 6
6 6
6 6
. ·7
Hajdu~ka. 6
6
Vaden"lea
Figure 46 Late Romanforts in the Djerdap

on the other sideof the Danubein that particularsection.It would be premature
to supposethat the Devil's Dyke in Sarmatiaand the Brazdalui Novac must be
of the samedate. But the evidencehitherto is in favour of a Constantiniandate
for the building programmeon the Upper Moesiansectionof the limes.
The fact that constructionwork on the Danubianlimes was not restrictedto
existing forts can be seenin the fort-lists of the Notitia Dignitatum, which con-
tain many new place-names.Nor is thereany lack of late Romanfortifications
revealedby excavationon the bank of the Danube,but their exact date of con-
struction has only occasionally been established.In default of sure dating
evidencea numberof fort-plans must be omitted from any reconstructionof the
history of the frontier.
Among Pannonianfortifications on the Danubeuncoveredso far (Fig. 47),
fresh buildings of the late Roman period occur in considerablenumbers.64
A common characteristicmust be pointed out: without exception they were
built on elevationsand, if the ground allowed it, on plateauxwith steepsides.
Further common featureswere the irregular outline normally adaptedto that
of the plateau,the small numberof gates(usually one only) and the greatthick-
ness of the walls comparedwith those of earlier forts. Thesecommon traits,
however,do not furnish any cluesfor dating, nor are the small finds, coins and
potterysufficient for closelyfixing the dateof construction.But thereis onetype
of tower which is certainly of Constantiniandate; it was preferredboth in new
buildings and also in reconstructionof earlier forts. This is the 'fan-shaped
tower'. It has severalvariantswhich were probablynot the result of typological
developmentbut local variationsof oneand the sametype. In new fortifications
thesetowers were built in such a way that their walls were radial extensionsof
the line of the fort-walls at the corner, or they were left open towards the
interior of the fort. 65 In fortifications which already existed the towers were
addedto the roundedcorner,and the curving part of the fort's cornerwas pre-
sumably pulled down in order to open the tower towards the interior.66 In
practically all forts on the limes a regularly recurring sequencecan be observed
at the angles(Fig. 18, p. 108); an internal trapezoidalcorner-towerdatingfrom
the first stoneperiod in or after the middle of the secondcentury; next, a small
roundedexternaltower dating from the end of the secondor the beginningof
the third century,and finally a long projectingfan-shapedtower of the time of
Constantine.The dating of the fan-shapedtower is basedon the following facts:
at the newly built, roughly triangularhill-top fort of Visegrad(PI. 14b) monetary
circulation starts only under Constantine;second, the renaming of Ulcisia
Castra as Castra Constantia67 is presumably connectedwith the rebuilding
associatedwith this type of corner-tower;and third, when the fort of Campona
282
Danube-
Budapest,
Budapest, Budapest, Esku tlr
Budapest,
oII' 50
iii:::'
100
metres
1
Hide61eloskereszt Tokod
Figure 47 Plansof late Romanforts in Pannonia
283
was destroyedin 322 and had to be reconstructedit was given fan-shaped
towers.
A further changemust be connectedchronologicallywith this type of tower.
So far, in at least three limes-forts, it has been establishedthat the portaeprin-
cipales or the porta decumanawere cut off from the outsideby a huge curving
wall (Fig. 18, p. 108); it hadthe effect of convertingthe former gateinto a single
complex of three towers.68 From now onwards the limes-road had to make a
detour, as can be seen particularly well at Camponain the behaviour of the
modern road. This alteration has a decidedly defensivecharacterand has its
parallel in the newly built fortifications with only one gate or two at the most.
The ditch was also retainedin the fourth century, and sometimesit was even
doubled.At someforts the ditch ran remarkablyfar from the walls, which had
the effect of providing a very wide berm betweenditch and wall. 69 The project-
ing towers therefore had the evident purposeof allowing an enfilade of the
besiegersin front of the walls.
The earthen ramparts which had supportedthe fort-wall from the inside
were sometimesremoved and buildings added to the inner face of the wall.
This can be demonstratedat the legionary fortress of Carnuntum70 or at Cam-
pona, and in all new fortifications the bank was omitted;71 presumablyit was
for that reasonthat they had thicker walls than the earlier forts of the limes.
Further types of towers, as yet not accuratelydatable,are semicircularpro-
jecting interval-towerswhich are characteristicof new late Roman building;72
horseshoe-shaped interval- and corner-towers 73 which, although sometimes
possibly of Constantiniandate, are probably in general characteristicof more

than one late Roman period; and lastly circular corner- and interval-towers
which probablycomelate in the history of late Romanconstruction:they occur
as subsequentadditions on late fortifications.74
In generalwe may concludethat the systemof strong-pointsand forts on the
ripa Sarmaticabecamesubstantiallydenserduring the fourth century. It is less
probable that a comparableamount of construction was going on on the
front of PannoniaPrima which faced the Germans.At present frontier re-
search-especiallyin the fairly recent excavationsof S. Soproni-hasshown
that the Danubebend in north-eastPannoniawas provided with exceptionally
strongdefences(Fig. 48). Whetherthe compactrow of fortifications is character-
istic of this section of the frontier only, or whether the results obtainedthere
may be generalizedto apply to the whole of the ripa Sarmatica,cannotyet be
decided.It is equally uncertainwhethercertain types of tower can be taken as
indicative of simultaneousbuilding 'activity along the whole of the bank of the
Danubeor only within shortersectionsof the frontier. The distribution of very
284
typical late tile-stampsbearing the namesof officers, units or places to some
extent supportsthe idea that the constructionof fortifications was differently
organizedin eachprovince: sometypesof stampare limited to the limes sector
of a single province; those of the limes in Moesia Prima and Dacia Ripensis
illustrate this particularly well.75 If the expansionof the letters DRP on a group
of tiles whose main distribution lies in the area of Dacia Ripensis,as D(acia)
R(i)P(ensis)is correct,76then we have in these tiles a terminus post quem for
extensivebuilding activity on this stretchof the bank, for Dacia Ripensisas the
nameof a province is attestedonly from Constantineonwards.
As canbe seenfrom walled-upgatesof forts andnewly-erectedhill-top strong-
holds, Constantinianbuilding activity on the Danubianlimes was basedon a
defensivestrategy.The representationof fort-gateson coin-reverses,frequently
repeatedsince the time of Diocletian, is proof that this defensivepolicy was
deliberatelypublishedas political propaganda. 77 Peaceon the Danube,of course,
had beensecurednot by thesefortifications but by the martial and diplomatic

successesof the emperor,and not least by the mutual animositiesof barbarian
neighbours.Nor is it likely that the title Sarmaticusassumedin 339 by Constans
referredto clashesof importanceon the Danube.Contemporaryliterature says
nothing of a war which could justify this title, and we might rather conclude
that totally unimportantfrontier skirmishesgavea welcomeopportunityto the
young emperorto award himself the title of Sarmaticusafter the exampleof
his father.78 Constansis, indeed,attestedin Pannoniain 339: in the springhe was
at Savaria, soon afterwards at Mursella north-east of Savaria, and only in
Septembercan he be shown to have beenat Naissus.79
After the death of Constantinethe Great, his sons assembledat Vimina-
cium80 to divide up the empire.The areaof the middle Danubefell to Constans,
who had beenplacedin the guardianshipof his eldestbrother, a guardianship
which he soon energetically rejected. Constansstayed in our area mainly at
Naissusuntil the spring of 340, that is until the defeatof the brotherwho was
his rival. Afterwards he twice again returnedto the Danubefrom the west: on
both occasionshe is attestedfor a short time at Sirmium (in 342 and 346).81
It was only in 350 that our provincesbecamethe sceneof important events
once again. Magnentiushad murderedConstansin 350 and had assumedthe
purple in the West; he was on the point of seizing the Danuberegion without
difficulty. It was the idea of Constantia,a daughterof Constantinethe Great,
to set up a counter-emperorin the Danubianarea,in order to savethe dynasty.
Vetranio, the old man chosenfor this purpose,was in no way a representative
of the Danubianarea. Although only a tool, he was remarkablyquickly pro-
claimed in Mursa and soon afterwardsin Sirmium, too ;82 this gives a hint of
28 5
the survival of certain old Illyrican tendencies.Vetranio was anything but
capable-presumably it was for that very reasonthat he had been selected-
and after deplorablehesitationshe finally moved out at the end of 350 to meet
ConstantiusII who was making a leisurely approachfrom the east;he greeted
him at Serdicaandon 25 Decemberabdicatedat Naissusat a ceremonyarranged
in as festive a way as can be imagined.83
ConstantiusII spent the winter presumablyat Naissus,or more accurately
not far from Naissusat the large villa owned by the dynastyat Mediana,84and
in the spring of 351 advancedto Sirmium where Gallus was createdCaesar.85
The following monthsbrought a heavy trial for ConstantiusII and once again
made southernPannoniathe sceneof war. Magnentiusappearedin Pannonia
from north Italy with a very large army, consistingin part of Rhenishbarbar-
ians.86 The whole of the summerwas taken up with a gradual retreatby Con-
stantiusalong the major road leading from Emonavia Poetovioand Mursa to
Sirmium. A first fight nearAtrans left Magnentiusthe victor. The latter did not
wish to pursue his enemy who in the meantimeproposednegotiations,but
advancedup the valley of the Savein the direction of Sirmium, took Sisciaand
attemptedthe siege of Sirmium. The army of Constantiusin the meantime
remainedat Cibalae. Apparently the adversarieshesitatedlong before they at
last joined battle south of Mursa on 28 September. 87 Magnentiushad to with-
draw to Italy and ConstantiusenteredSirmium where he stayedtill May 352.
At the battle of Mursa the empire suffered an enormousloss of men fit to
bear arms. Neverthelessthe situation was not exploited by the barbariansnear
the frontier of Pannonia;along the border quiet reigned, a further proof that
the wars with the barbarianscannot be explained by the simple formula of
Roman-barbarianconfrontation.Since the developmentof the client-systemin
the first century, great attacksby the barbarianswere causedby the situation
beyond the frontier rather than that within the empire, and the wars that
ConstantiusII had to wage on the Danubea few years after his victory over
Magnentiuscan also be interpretedonly in the light of this consideration.
In 356 whilst the emperorwas in Rome, the Quadi, who had not beenheard
of for many decades,overran the province of Valeria, while the Sarmatians
attackedPannoniaSecundaand Moesia Prima, that is the limes on the southern
borderof the plain.88 How far theseinvasionspenetratedinto the provinceswe
do not know;89 it by no meansfollows from the later negotiationsfor the return
of prisonersof war90 that the invasionswere comparableto the great disasters
under Marcus or Gallienus. The brevity with which Ammianus Marcellinus
dealt with this attack suggeststhat it was itself of no greatimportance,though
its consequences for the future on the ripa Sarmaticawere far-reaching.
286
ConstantiusII arrived in Pannoniain the summerof 357 and enterednegotia-
tions with the Quadi and Sarmatians. 91 The following yearswere takenup with
various operationsand negotiationson the ripa Sarmatica,the emperormean-

while staying continuouslyin Sirmium. He left Sirmium only in the middle of
359, to go to Constantinopleon the news of the fall of Amida.92 Our only
source, which is however very detailed and occasionally well-informed, is
Ammianus Marcellinus, though his record contains a few inconsistencies. 93
For instance,the information that the Quadi and Sarmatianshad attackedthe
empire during the emperor'sstay in Rome can presumablybe taken to be a
duplication, since the sameis reportedfor the winter 357-8 in similar words,
and according to Ammianus the emperor travelled from Rome to Illyricum,
whereasno tracesof sucha journeyare to be found in the well-attesteditinerary
of Constantius.It was in the summer of 357 that Constantiusfirst came to
Pannonia.
The fact that the emperor,though presentin Sirmium in 357 and receiving
emissariesfrom the Quadi and Sarmatians,did not start his great expedition
before 358, suggeststhat the situation in 356-8 resembleda bellum suspensum;
the barbariansmade demandswhich were then the subject of negotiations.
Only the attackof the Quadi and Sarmatiansin the winter of 357-8 inducedthe
Romansto interrupt thesenegotiationswhich were being carriedon within the
setting of the presumablyvery complicatedclient-system,and to proceedto
war. The report of Ammianus on the expedition of ConstantiusII is written
very muchin the vein of the late Romanideal of the triumphantemperor.There-
fore we learn details only from the homagespaid by defeatedpeoplesand little
is said aboutwhich of the many small kingdomswere in fact defeated,which of
them were merely intimidated, and where the fighting actually took place.
Ammianusgives a fairly long list of Sarmatianand Quadankings, eachof whom
apparentlypossessed an alliancewith Rome of his own and who were indepen-
dent of eachother. They arelisted, however,only after the emperor'svictorious
advance,whereasthe descriptionof the expedition,which is in any caseshort,
merely talks of Sarmatiansin general.
The expeditionstartedin April of 358 when ConstantiusII led his army from
PannoniaSecundaover a pontoon bridge against the Sarmatians.Evidently
great storewas set by the Romanattackachievingsurprise.After their country
had beenplunderedfar and wide, the Sarmatiansretreatedinto valleys which
can only have lain among the Dacian mountains on the easternedge of the
Lowlands. From this it can be deducedthat the expedition advancedtowards
the north-east.Another army advancedfrom Valeria and looted and burnt in a
similar fashionin the northernsectorof the Lowlands.The emperor'sarmy then
287
effected a junction with these other troops and the united army invaded the
territory of the Quadi.
This is the extent of our knowledgeabout the expedition. Ammianus con-
tinues his report with the negotiationsfor peaceand with the homagespaid.
First Zizais, a youth of royal birth, approachedthe emperor, and then,
encouragedby the mild treatment received by Zizais, came Runo, Zinafer
and Fragiledus.Among other things they all proposedto put themselvesand
their people under Roman rule: in other words they offered themselvesfor
receptio. After a client-treaty had been made with these chieftains, Araharius
and Usafer, kings of north-westerlySarmatianand Quadanpeoples,madetheir
appearance.
A particularproblemwas posedby the Argaraganteswho appearedonly after
the others.As we haveseen,thesehad oncefled to the Victohali. They believed
that now their time had come. They came forward with the demandfor a
guaranteedfreedom. They were given Zizais as their king.
All these negotiations were carried out by the emperor in person.
In Ammianus'mannerof description,they havethe appearance of a continuous
ceremonyin conqueredcountry rather than peace-negotiationswith peoples
all equally defeated.The possibility cannotbe excludedthat the imperial expedi-
tion in the spring was nothing but a demonstrationfor the purposeof intimida-
tion, in order once againto convinceneighboursof the frontier underpressure
from beyondthat Rome intendedneither to conquerthem nor to acceptthem
into the empire.
The negotiationswith the Sarmatianstook place somewherein the country
beyondValeria. From here, court and army proceededto Brigetio, the baseof
one of the expeditions into the country of the Quadi. The Quadankings-
Vitrodurus, Agilimundus and others-werereadyat onceto makea new client-
treaty with the Romans.
As we have seen,the Romanswere only interestedin maintainingthe system
of client-kingdoms along the border. The first king of the Sarmatianswith
whom negotiationswere enteredwas Zizais, who was instituted as king of the
Argaragantes.From this we may suggestthat in the last resort the unrest had
startedwith the Argaragantes,the problem of whom had not been solved in a
sufficiently radical mannerby Constantine.In the war of 332-4 Constantinehad
beensatisfiedwith receivinginto the empiresomeof the Argaraganteswho had
been expelled by the Limigantes, and letting the rest of them disappearfrom
Romansight, and he laid greaterstore on obtaining control of the Limigantes,
the new mastersof the lands beyond the borders of southernPannoniaand
Moesia. Although the exiles-exsulespopuli in Ammianus-foundhospitality
288
amongthe Victohali about 334, they probably bore the main responsibilityfor
the ultimate assumptionof arms by the Quadi.
That the initiative for the Sarmatian-Quadanattackin 3 5 and
6 3 5 came
7 from
the Argaragantesalso follows from a phraseusedby Ammianus.9 4 The expedi-
tion of ConstantiusII was thus aimed at peoplesmade restlessby the Argara-
gantes.The fact also enablesus to understandthe demandof the Argaragantes
that the Romans should grant them their freedom, for they had had bitter
experiencesfrom the armed assistancegiven by Rome in 332: Constantine
though supportingthe Sarmatiansagainstthe Goths had not reconstructedthe
old-accustomedsocial stratification.
The attitudeof the Limigantesduring the expeditionagainstthe Argaragantes
and the Quadiis very obscurelydescribedby Ammianus.It is statedthat theyhad
felt inducedby the activity of the Argaragantesto attack the Roman frontier,
compelling ConstantiusII to punishthem as well in the end. Sincethe Danube
borderof Moesialay oppositethe Limigantesand sinceaccordingto Ammianus
this sectionof the frontier had also beenattackedin the winter of 357-8, it is
probablethat the Limigantesdid indeedundertakean action of their own. But
we cannotexcludethe possibility that the new treatieswith the other Sarmatians
had arrangedfor commonRoman-Sarmatian action againstthe Limigantes:for
it was Zizais in particular who took part as a Roman ally in the campaign
againstthe Limigantes,and accordingto Ammianusthe action was aimedfrom
the beginning at the resettlementof the Limigantes. In his summary of the
outcome of the war, Ammianus stressestwo results: the restorationof their
land to the peoplewho had beenexpelled,and the establishmentof a new king,
that is one recognizedby treaty as a Romanclient. Thus one can perhapscount
the solution of the problem of the Limigantes among the main aims of the
expedition. Ammianus devotes more attention to this part of the expedition
than to the first part. In the campaignagainstthe Limigantes,the Romans,the
rest of the Sarmatiansunder Zizais, and the Taifali proceededin alliance. The
revolt of 332 was revengedwith unprecedented cruelty, but it was not possible
to reconstitutethe old social order. The Limiganteshadto evacuatethe country
oppositePannoniaSecundaandMoesiaPrima, which they had occupied,and in
it the Argaraganteswere permitted to settle again. The Limigantes received
land elsewhereand also a client-treaty.
After these achievementsConstantiusII victoriously entered Sirmium in
June 358. There he spent the winter, during which news arrived that the
Limigantes had left the land assignedto them and that there was danger of
disorder once more in Sarmatianterritory owing to their raids. In April 359
the emperorthereforeagain preparedan expedition. This time he moved with
289
his army into the province of Valeria and from there sent two tribunes to the
Limigantesto call them to account.The Limigantesansweredthat they desired
to be acceptedinto the empire and to be given land in some part of the orbis
Romanus.Constantiusconsideredthis a conclusivesolution to the problem of
the Limigantesand preparedfor a ceremonyof which the scenechosenwas the
bank of the Danubenot far from Aquincum. The only precautionto be taken
was the placing in readinessof a unit of the fleet on the Danube. The Limi-
gantes, however, on being admitted to the emperor'spresenceattempteda
surpriseattack on him; in the ensuingtumult Constantiusbarely escapedwith
his life, and the Sarmatianswere only put down after a hard struggle. But the
massacresecuredquiet on the ripa Sarmaticaand peacelastedtill 365.
ConstantiusII soon afterwardsleft Pannoniafor his last visit to Constantin-
ople and the East.95 The changeof governmentin 361 affectedour provinces
only inasmuchas Julian, descendingthe Danubein secret,had to makea short
stop at Sirmium and afterwardsstayedin Naissustill the end of November,
where the news of the death of Constantiusreachedhim.96 In the autumn of
this year massingof troops did take place in southernPannoniaand on the
diagonal road acrossthe Balkans;97 but on the whole the changeof govern-
ment was peacefuland there are a few indications98 that paganismin Illyricum
was still strongenoughto give occasionalexpressionto loyalty towardsJulian.
On the occasionof the proclamationof the Moesian Jovian and of the Pan-
nonianValentinian,an Illyrican junta of the old sort hardly playedany decisive
part. The Pannonianorigins of the Valentinianic dynastygave rise to repeated
controversiesbetweenIllyrian and non-Illyrian cliques-someof this can be
seenin the attitudeof the Historia AugustatowardsSirmium andthe Pannonians,
and Ammianusalso was distinctly anti-Pannonian-but Valentinianhimselfwas
anythingbut representativeof an Illyrican trend. He did not residein Sirmium,
and on his visit to Pannoniain 375 he showedlittle feeling for the complaintsof
his fellow-countrymen.99 It cannot,of course,be deniedthatthe choicefell on the
Danubianofficers JovianandValentinianbecausethe officers makingit hadsome
awarenessof the prestigeof the old Illyriciani. Both emperors,who successively
enteredinto the heritageof the Constantiniandynasty,camefrom the surround-
ings of Sirmium: Jovianfrom Singidunum100 and Valentinianfrom Cibalae.101
They presumablyowed their military careerssimply to the circumstancethat
their family homeslay in the vicinity of animperial residence.A numberof other
important personalitiesof this period were Pannonianstoo, and they obtained
swift promotion for the reasonpresumablythat they were fellow-countrymen
of the emperors.But in the secondhalf of the fourth century Sirmium had
alreadylost its central role. In the division of the empire betweenValentinian
29°
and Valens, Sirmium happenedto fall in the periphery of the two areas;the
centresof gravity of eastand west then becamefinally fixed, and the emperors
stayedin the new centres.
Valens and Valentinianhad sharedcourt, power and empire at Mediana,the
large imperial villa outside Naissus.102 In June 364 they were still attestedat
Naissus,and in the summerthey were both in Sirmium.103 Next year Quadan
and Sarmatianattacks are reported,I0 4 which were probably dealt with by the
receptio of some Sarmatiangroup; for not long afterwards the Gallic poet
Ausoniuswrites of the 'recentlysettledSarmatians'in the valley of the Mose1.I05
The barbarianattacks-whichare attestedon practically all frontiers of the
empireI06-were presumably connected with the change of government.
Possiblya parallel can beseenin the 'Sarmatianwar' of Constans,which broke
out soon after the deathof Constantine.It may have beenmore or less regular
for barbariansbeyondthe frontier to considertheir treaty as void on the death
of the emperor,or to fear that the new emperorwould make a less favourable
treaty. However, the death of an emperorcould also give rise to the hope of
a more favourabletreaty.
The attack of the Quadi and Sarmatiansin 365 could be dealt with easily and
briefly. The centreof unrestshifted into the areaof the lower Danube,the ripa
Gotica, where fortification building was started energeticallythat year.107 But
building activity on the limes, which receivesgood noticesin the contemporary
literature, was not causedby the immediatesituation, for it was soon taken up
on the ripa Sarmaticaand on the limes of PannoniaPrima as well. It is one of the
most seriousproblemsof Pannonianfrontier researchthat this activity, which
was probably extensive,cannotas yet be sufficiently clearly distinguishedfrom
the other periods of construction,which are themselvesonly roughly deter-
minable. For a long time an apparentlyreliable clue was providedby tiles bear-
ing the stampsof the ducesFrigeridusand Terentianus-both in Valeria-datable
to the year 375. And since thesetiles are mostly found associatedwith other
late Roman tiles, many other tile-stampswere consideredto be also certainly
of Valentinianic date. But it has recently been shown that some stamps-
especiallythoseof the so-calledOFAR group, which are attestedin the fortified
landing-placesof the ripa 5armatica-cannotbe of Valentinianic date, and that
the dux Frigeridus cannothave functioned in Valeria in 375.108 The Valentin-
ianic period of constructionon the limes, which is certainlyimportant,therefore
cannotbe clearly defined. The most important questionis still undecided:did
Valentinian mainly repair existing forts and, apartfrom that only order a great
numberof burgi to be built, or are the fairly numerousnew late Romanforti-
fications attributableto him? If so, this would meanthat a colossalamountof
291
The Danubejrontier in the late Romanperiod
Watch
Watch -tower
-tower
Watch -tower
Fort
,Oil' . .S' km
Figure 48 Late Romanforts on the Danubebend
work on the limes was carried out by Valentinian, A fortunate discoverymade

recently shows that the very numeroustiles with the stamp Frigeridus vir per-
jectissimusdux, known from many placesin Valeria, must be datableto 372 at
the latest,I09The tiles in question,however,were roof-tiles, which attesteither
the completionof the buildings from which they came,or elsemerereconstruc-
tions and repairs:they do not prove new buildings, The only certainnew build-
ings of Valentinianic date are a few burgi and one fort, which are named on
building-inscriptions,uoBut these very inscriptions give rise to the suspicion
that Valentinian orderednew constructionat only a smannumberof places:it
is certainly curious that small watch-towers,never exceeding8 by 8 m in size,
were given a long, pompous inscription. The problem becomeseven more
complicatedwhen we take the distribution of stampsof the Frigeridus group
and the find-spots of the building-inscriptionsinto considerationas well: prac-
tically all come from the Danubebend, where recent excavationshave indeed
discoveredan astonishingnumberof burgi, all of the samesize (Fig. 48).III Does
this indicate thatValentinianicbuilding activity was restricted_tothe sectionof
the limes aboveAquincum and below Brigetio?
The director of constructionwas Aequitius, magister militum, named on all
Pannonianbuilding-inscriptions.Work probably startedsoon after the acces-
sion of Valentinian, as the constructionof a 'fort with towers' was completed
in 367 if not earlier. Two burgus-inscriptionsare datedto 37! and 372 respect-
292
ively; the latter was recently found in a small btlrgus at Visegrad(PI. 36b). The
other btlrgtls was built for the supervisionof trade acrossthe frontier. Unfor-
tunately the find-spot of the building-inscription relating to the fort is not
known accurately;it canbe relatedto one of severalforts in the vicinity of Solva
(Esztergom).
Buildings on other sections of the Danube frontier are even less easily
identifiable. A building-inscription-unfortunatelyfragmentary-cameto light
in Carnuntum;II2 officer-stampsof the Frigeridusgroup are met with in Valeria
in other places too, and small btlrgi, correspondingin every detail to that of
Visegrad,are provedin someplaces,thoughneverso numerousas on the bend
of the Danube,where the hilly bank and the meanderingof the river (PI. I4b)
necessitateda compactrow of towers.II3
By 373 building activity was extended to the land beyond the frontier.
Ammianus reports that Aequitius had a fort erectedon Quadanterritory,II4
and a lucky find has recently revealed the presenceof a btlrgtls of Valentin-
ianic type in the countryof the Sarmatians.IIS This btlrgtls stoodon the northern
border of the Lowland, about 60 km away from the Danube, and so deep
inside Sarmatianterritory that the security of the small garrison could only
havebeenguaranteedby treaty and/or by the presenceof Romantroopsin Sar-
matia. There is further evidencefor the stationingof Roman soldiers there in
the Notitia Dignitattlm, which lists the two forts built in Sarmatiain 294 as
lying in barbarico.II6 They werethus still occupiedin the secondhalf of the fourth
century.Thereis no mentionof Romanpostsin the countryof the Quadibefore
Valentinian,and from Ammianus'descriptionwe may be surethat his construc-
tion of forts there was indeeda new experiencefor the Quadi, sincethe events
of the years374-5 were caused by this very work. BeforeValentinianthe Quadi
hadnot beendangerous,as-amongstotherevidence-theresultsof fighting on
the approachesto Pannoniatestify. Throughoutthe fourth centurythey always
appearmerely as easily pacifiable allies of the Sarmatiansand were never of
more than secondaryimportanceon the Pannonianlimes, as is also emphasized
by Ammianus.II7Valentinian'sdesireto occupyQuadanlike Sarmatianterritory
with watch-towersand garrisonsis either basedon a new strategicconception
or is due to increasingtension in the Carpathianbasin. Whether one should
attach importanceto Ammianus' expression:'as if the country of the Quadi
were already under Roman rule',II8 I would prefer to leave undecided.It is,
however, strangethat Ammianus speaksof only one fort in the country of the
Quadi, and that Valentinian'spunitive expeditionof 375 did not advanceinto
the centre of the country but attackedtheir easternborder, the hilly region
oppositeAquincum. It follows from the negotiationsof ConstantiusII in 358
293
that the centreof Quadanterritory was near Brigetio. It is thereforequite pos-
sible that the fortification was in fact a link in the chain of occupation of
Sarmatiancountry, and was to be constructedsomewherein the Sarmatian-
Quadanborderarea. Since Roman posts could not be erectedin a barbarian
country without a previous war or treaty we must assumethat the Quadi
lodged a protestwith Aequitius, on the ground that the fortification had been
started without a previous Roman-Quadanagreement;and that the reason
why it had beenundertakenwithout this previous agreementwas becausethe
place in question was in a border area, presumablythe northern border of
Sarmatia.
This interpretationof eventsis confirmed by the fact that Aequitius imme-
diately ordered construction of the fortification to be stopped, when the
Quadanemissarieshad protestedagainstit. Becauseof the delay in its building,
Aequitius sufferedan accusation,and the youngMarcellianus,who was the son
of Maximinus from Sopianaein Valeria, was made dux Valeriae. The young
man orderedthe constructionof the fortification to be resumed,and committed
the catastrophicblunderof inviting Gabinius,the Quadanking, underthe pre-
text of negotiations,and then killing him.
The Quadi decidedto avengethis outrage.This time they were the ones to
call in the armed assistanceof the Sarmatians,and both peoplesinvaded Pan-
nonia in July 374. The invasion found the provincial populationentirely un-
prepared;the country peopleengagedin harvestwere slaughteredor dragged
off, and the daughterof ConstantiusII, on the way to join her bridegroom
Gratian, was almost taken prisoner. From the large number of coin-hoards
endingwith coins of ValentinianI 19 from all partsof the province,variousinva-
sions,by severaldifferent groups,can be inferred, thoughnew study showsthat
somefinds which weretraditionally associatedwith this war mustbe excluded.I2°
The imperial city of Sirmium was in greaterdangerthan it had beenfor more
than a century. Probusthe praetorianprefecthad the neglectedfortifications of
the town repaired:money from the city treasury,which had beencollectedfor
building a theatre, was made use of for this purpose. Two legions, which
togethercould have got the upper hand againstthe Sarmatians,were not co-
ordinatedbecauseof staff rivalries and were defeated.Only the dux of Moesia
Prima, the later emperor Theodosius,succeededin defeating the plundering
barbarians.They retreatedand declared themselves ready to comply with the
demandsput to them: they were all the more readyto do so, sincenews of the
arrival of troops from Gaul diminishedtheir chances.
So far the accountof Ammianus: he also gives a hint that it was the Sar-
matians who did the larger part of the damagein Pannoniaeven though the
294
Quadihadbegunthe war. WhenValentinianleft Trier for Pannoniain the spring
of the next year, Sarmatianemissariesdid in fact come to meet him, with a
variety of promises.
Valentinianarrived at Carnuntumaboutthe end of April or the beginningof
May.121 He was accompaniedby the imperial family with the court and army
and establishedhis headquartersnear Carnuntum.From information given by
Ammianus it can be deducedthat the court lay at the Villa Murocincta, 100
Romanmiles from Brigetio and in the vicinity of 'desertedand ruined' Carnun-
tum, i.e. probably in the huge villa of Parndorf.I22 The fact that Valentinian
had to remain there for three months-thewhole of the summer-isevidence
that the purposeof his visit was not merelythe preparationandcarryingthrough
of an expedition,but that a numberof affairs requiredputting in order in Pan-
nonia also. Even the preparationsfor war, of course,met with difficulties: since
the war of the previous year had happenedduring harvest,provincial stocks
were presumablynot sufficient for a large expeditionaryforce. The tax policy
of Probus, the praetorian prefect of Illyricum, moreover, had given rise to
generaldiscontent,giving the emperormuch to do.
At the end of the summerthe Romanarmy at last departedand attackedthe
Quadi from Aquincum. A pontoon bridge was built quickly and unnoticed
north of Aquincum, and the imperial troops surprisedthe Quadi who did not
dareto fight in the open. The cavalry underMerobaudeshad alreadystartedan
advancefrom a different point. Valentinianplunderedand burnt the country of
the Quadi, and at the beginningof Novemberreturnedto Aquincum. Quadan
emissarieswerereceivedat Brigetio, but during this receptionthe last Pannonian
to occupy theimperial throne was overtakenby deathon 17 November.
The next dayswere occupiedin Brigetio with negotiationsinside the supreme
command,since it was to be feared that the army would demandan emperor.
In order that an imperial election might be conductedin the campin the least
damagingmanner, the boy Valentinian II was fetched from Murocincta and
presentedto the soldiers. The proclamationfell on 22 November,whereupon
mother and son went to Sirmium, which was chosen for them as their
residence.123 ValentinianII stayedin Sirmium until 378; he was the last emperor
to residein this famous Pannoniantown.
How the problem of the Quadi and Sarmatianswas solvedwe do not know.
Presumablythe client-treaties had to be renewed, and possibly the annual
paymentsincreased.At any rate we hearnothing during the following yearsof
any difficulties from the Quadi or Sarmatianson the Pannonianlimes. But on
the lower Danubethe situationbecameall the more aggravated;here,pressure
by the Huns had causedmovementsof the various Gothic groups. In 376 the
295
receptiowas,in fact, given to someof the Gothic princes,andthis causeda chain-
reaction of catastrophicevents. These incidents in the Balkan peninsula, in
which the areaof Upper Moesiawas also affected,openeda new period charac-
terized by the settlementoffoederatiin the Danubearea.
The middle Balkan area,that is the provincesof MoesiaPrima, Dacia Ripen-
sis, Dacia Mediterraneaand Dardania, had enjoyed a peacewhich had lasted
from the Tetrarchyuntil the deathof Valentinian; its peoplesowedit aboveall
to their geographicalsituation. The frontier district of Moesia Prima formed
part of the ripa Sarmatica,the centreof which was in Pannonia.The frontier of
Dacia Ripensiswas the westernpart of the ripa Gotica-wherebarbarianpres-
sureon the Lower Moesianlimes was concentrated.The middle Balkan areawas
of importanceonly as the deploymentzone betweenthe Sarmatianand Gothic
fronts, and becamethe sceneof fighting only when barbariantroops stood on
Romansoil. When once the Goths had crossedthe frontier of Lower Moesia,
however, the centre of war shifted to the Balkan peninsula,and the strategic
importanceof Upper Moesia showed up clearly. As deploymentzone it was
usedmuchlessfrequentlyfor the settlementoffoederatithan were the remaining
Danubianprovinces,but from the end of the fourth century onwardsit had as
much to suffer as its neighbouringterritories.
Pressureby the Huns was thus at first advantageous to Pannonia,whereafter
the Gothic invasion of the Balkans the crisis was relaxed. The Sarmatiansand
Quadi were presumablyeasyto appeaseafter 375, becauseGothic pressureon
the Lowland had decreasedas the result of the Balkan invasion.
296
Chapter 9
The final period of prosperity
The late antique period in the Danubian provinces is often characterizedby

historiansas one of gradual decline and disintegration.Responsibilityfor this
rather one-sidedgeneralizationmust be laid partly on the sourcesthemselves
andpartly on the limitations of archaeologicalresearch.As far as the sourcesare
concerned,it is often pointed out that late antique and early Christian inscrip-
tions are rarities in the provinces;I gravestoneswere rarely erectedany more,
andthe absenceof building- andvotive inscriptionsis due simply to the poverty
of the formerly rich and votive-loving municipal aristocracy.Stone sculptures
becameso rare that one might doubt the continuedexistenceof stonemasons'
workshops.As far as archaeologicalresearchis concerned,it long restricted
itself, particularlyin our provinces,to the limes and to the towns on the bank of
the Danube,whereit is true that appreciableprosperitycannotbe proved. It is
only recentlythat doubtshave beenexpressedwhetherit is legitimateto extend
the symptomsof decay, which undoubtedlyexistedalong the Danube, to the
interior of the countryandto all socialclassesas well. The declineof inscriptions
and the poverty in the frontier area are in contrast to a number of written
sources,which report large and flourishing towns in Pannoniaand Moesia,
and count our provinces among those which were wealthy or at least self-
supporting.2 The excavationsof the last decadesin certainof the towns furnish
unequivocalproofs that no declinecan be seenat leastin the towns in question.
A careful considerationof the decline of Roman civilization in the Danube
provincesis thereforeopportune.
As we havementioned(pp. 266, 272), severalimperial measurestowardsthe
end of the third century indicate an intention to make agriculturein Pannonia
297
The final period oj prosperity
and Upper Moesiaprofitable. The EmperorProbusorderedvines to be planted
at the Alma Mons(FruskaGora) north of Sirmium andat the AureusMons eastof
Singidunum,and theseplantationshe 'put at the disposal of the population';
furthermore,it is reportedthat he orderedthe marshyareaeastof Sirmium to be
drained'in order to rendernative farmlandfertile and to increase itsextent'. A
road-stationeastof Sirmium carried the nameFossis,3and the modernSerbian
place-nameis Jarak,which means'ditch'. Improvementswerealso carriedout by
Galerius in the province of Valeria by canalizing the natural outlet of Lake
Pelsoin order to gain arableland. Theseworks, orderedon the highestlevel,
must have been consideredof great importance at the time, and therefore
they have been stressedin the concise imperial histories of the late fourth
century.4
At the sametime privatepersonswerealso presumably calleduponto produce
moneyand man-powerfor this 'economicplan': a curious altar from the north-
west corner of PannoniaSecundais intelligible only againstthe backgroundof
this economicpolicy enforcedfrom above. This altar was dedicatedto Liber
Paterby a certainAurelius Constantiusandhis sonVenantiuson the occasionof
vines being plantedon a very sizeablepieceof land 400 arpennesin area(= 200
iugera = c. 100 acres).The kinds of wine are enumeratedon the altar: Cupenis,
Terminis,Vallesibus,Caballiori. Unfortunatelythesenamescreategreatdifficul-
ties. But it is interestingthat the size of the piece of land is expressedin Gallic
units of area.Whethergroundin Pannoniawas generallymeasuredin arpennes
we do not know, and we are therefore not able to decide whether Aurelius
Constantiusmay havecomefrom Gaul. The recordingof a vine-plantingon an
altar we must supposewas due to the importanceattachedto this activity at the
time.s The dating of the altar does not offer much difficulty, for the names
Constantiusand Venantius are typically late antique though the pagan vow
suggestsnot too late a date.Attention shouldbe drawn to the circumstancethat
the nameandrank of Aurelius Constantiusareerased;evidentlytheestate-owner
fell into disgraceandit will be permissibleto draw the conclusionthat his estate
passedinto imperial possession.And this conclusionbrings us to the period of
internal unrest under the later soldier-emperorsor to the later years of the
secondTetrarchy.
Not far from the find-spot of this altar a mountainis attestedin the Itineraries
calledAureusMons-whichis not to be confusedwith the placeof the samename
in UpperMoesia6 -wherethe vineyardsof Constantiuscan be lookedfor. This
mountain,like the Alma Mons and Aureus Mons in Upper Moesia which had
beenplantedwith vines by Probus,lay on the bank of the Danube;the river
providedthe transportof the wine. Viticulture in Pannoniais attestedby finds
298
suchas vinedressers'knives,7which are also representedin the handof the deity
Silvanus (PI. 3I a), and by a few wine-presses;8 Dio talks of Pannonianwine,
however,in a very disparagingmanner,9thoughat the time the surroundingsof
Aquileia are praisedby Herodianas the wine-gardenof Illyricum. 1O Pannonia
was thus dependenton imports of wine evenin the Severanperiod, andwe have
to take the vine-plantingsby Probusas a measureaimed at securingeconomic
independencefor the province. Nevertheless,as late as 383 it is reportedthat
Pannoniasupplied corn to north Italy and was given wine in exchange. II It
would be very interestingto learnwhat effect the vine-plantationson the Danube

hadon the trade-balanceof PannoniaandMoesia,andhow north Italy reactedto
them. It is, of course,quite possiblethat the vineyardsin PannoniaSecundaand
Moesia Prima mainly servedto supply the populationof Sirmium.
Villas and their estatesin the fourth century

The plantations and improvements towards the end of the third century
presumablywent handin handwith the growth of largeestates.In somepartsof
PannoniaandMoesialarge estatesmadetheir first appearanceat that time. This
processled to Pannoniasoonbeingcountedamongtheprovincesexportingcorn
and cattle. Late antiquesourcesdescribePannoniaas a rich and cheerfulland.12
The Expositio totius mundi (c. 350), which attests the agricultural exports of
Pannonia,also mentions that Dardania supplied Macedoniawith cheeseand
lard, and gives the information that Moesia was only capableof supporting
itself. So Dardanianhill-farming had become capableof exporting, and the
agricultureof Moesia too enjoyeda comparablegrowth.
The predominanttype of estatein the late Romanperiod seemsto have been
of mediumand of large size. The estatescan be assessed with someprobability
from the size of the villas. The large villas of former periodscontinuedin usein
the fourth century; they were evenenlargedand furnishedwith mosaics,as for
instancethe villa of Deutschkreuznear Leitha13 and in the sameareathe huge
villa of Parndorf,to be discussedbelow. It is, however,much moreimportant
that we already have a very large number of villas and of partly explored
dwelling-houseswhich were first built in the fourth century.I4 Some of them
were decidedly luxurious, as shown by stucco decoration,architecturalstone
fragmentsand wall-paintings.Often the bath was in a building of its own, and
this also suggeststhat the villas werelarge.Is In areaswhereevenin earliertimes
largeestateshadbeenthe predominantform of land-ownership,suchas north of
Lake Pelso,I6 new villas were erected,but large and small villas can also be
demonstrated in areaswheresimilar buildings arenot previouslyattested.In this
299
contextthe provinceof Valeria requiresspecialmention:an imposingnumberof
late Romanvillas have beendiscoveredwithin its borders.
As far as size and groundplan of late villas are concerned,we find a variety
basedon the introduction of ground plans which had not hitherto been em-
ployed in the province (Fig. 49)' Apart from villas with a centralperistyle or a
central room,I7 there are large, sometimeseven huge, palaceswith corner-
towers or wings,18 other buildings erected around a large rectangularinner
courtyard,19and also a number of smaller dwelling-houseswhich mayor may
not belong to an agricultural estate.20 Thesesmaller dwelling-housesare distri-
butedparticularly thickly in the near vicinity of Aquincum,21but this doesnot
imply that elsewhereonly bigger villas were built.
The exactdating of the late villas and rural dwelling-housesis as uncertainas
that of the late Romanlimes. For a few a date in the Constantinianperiod has
beenproposed,but it is not impossiblethat otherswere erectedevenlater. The
processof concentrationof estatestherefore probablycontinued during the
fourth century;in any casewe haveevidencethat Pannonianestate-owners were
living on their estatestill the end of that century.22
Within what was formerly Upper Moesia we have so far only a few ground
plansof villas, for instancethat of Remesiana,which belongsto the type of huge
estate-centrewith a large central court;23 a palacefrom Kostol near Zajecar24
lying in the areaof the supposedmunicipium Aurelianum; and amongothersa
dwelling-housewith bath-suite25 in the mining district on Mount Kosmaj. The
date of constructionof thesevillas and palacesis not known, but it is striking
that from this areawe havea few hoardsandisolatedfinds-of the fourth century
which unequivocallypoint to extremelyrich, high-rankingowners.26 The state
cameofrom Kusadak,27south-eastof Singidunum,was certainly in the posses-
sion of a fourth-century man of high rank, and the silver hoards including
silver dishes(PI. 35 b) commemoratingthe decennaliaof Licinius 28 (p. 277) attest
rich people, too. We know much more of the imperial estatesand palacesof
UpperMoesia,aboutwhich the evidenceis availablein our sourcesandfrom the
results of excavation.
Among the villas were some that belonged to imperial estates,and were
furnished as palaces.It is a priori probablethat Pannonianestate-owners,who
were involved in the political strugglesof the military anarchyand Tetrarchy,
did not always emerge as members of the successfulparty, but often lost
influence, wealth and even life, and suffered the confiscation of their estates.
Moreover, emperorswho originated in the Danubearea presumablydid not
omit to erecta palaceor summerresidenceat their birthplace;thesepalaceswere,
of course, at the same time centres of large imperial estates round about.
300
GOR.SIVM Tac
H05S1.uheteny
o 10 ~o
, 1 I
metres
5umeg
Komlo
REMESIANA Bela Palanka
I
Csucshegy
AVRELlANVM?
Kostol
Csucshegy Csucshegy II
Csucshegy
Csucshegy Budakalasz Testve.rhegy

o 50 100
..
' II1II=' metres Graves
Figure 49 Plans of late Romanvillas and houses

The final period oj prosperi!),
Maximianus Herculius orderedthe erectionof a palaceat his birthplacenot far
from Sirmium,29Galeriushada palacein his birthplacein DaciaRipensis,named
Romulianaafter his motherRomula,30and the Constantiniandynastyownedan
immenseseatat Mediana,31a few miles eastof Naissus,whereit is possiblethat
Constantinethe Great was born; it is presumablyjustifiable to maintain that
influential representativesof the Illyriciani at the imperial court had luxurious
country-housesconstructedfor themselvesin their homeland.
Of the imperial palacesmentioned,that of Mediana has already been partly
investigated.3 2 The portion uncoveredconsists of state rooms provided with
mosaicfloors (PI. 43b), walls encasedin marble,and rich plastic decoration.Not
far awayhouseholdroomswith hugestorage-containers andwine- or oil-presses
came to light. The area enclosedby the wall of the Pannonianvilla of Parn-
dorf33 was almost as large as a legionary fortress, and behind the palacelong
barrack-likebuildings werefound (Fig. 32, p. 170). At Parndorfit is not only the
size of the site which suggestsan imperial owner, but evenmore so the greataula
or audiencehall added about the end of the third century, and the luxurious
furnishing of the palace: almost all the larger rooms were laid with mosaics
aboutthe beginningof the fourth century.It is not thereforerashto follow Saria
in the belief that this villa was the scene of the conferenceof emperorsat
Carnuntumin 308. It is evenmore likely that the family and court of Valentinian
stayed in this villa in 375: the summer-residenceof Valentinian, the villa
Murocincta, lay after all in the neighbourhoodof Carnuntum,100 miles away
from Brigetio. All this fits well with Parndorf.
Therewere of courseimperial palacesin the towns too, but they often lay in
the suburbs,not in the centreof the town. A palacefurnishedwith mosaicsnorth
of Poetovio was excavatedin the nineteenthcentury; it too is mentionedby
Ammianus.34 He also writes of the regia at Savaria,35but we may expect the
largest residenceat Sirmium. The palaces of Poetovio, Parndorf and a few
others to which we shall return later were equippedwith mosaics which in
ornament,style andexecutionarestrikingly uniform andwhich canbe attributed
to a centralworkshop,possiblyevento the sameschoolthat producedthe early
Christian mosaicsin Aquileia.36
Within the rectangularwalled enclosure of the villa of Parndorf a huge
horreum (store-building)was uncovered(Fig. 32), proving that Murocincta was
the centreof a large estate.This brings us to a questionwhich so far has found
no answer,but whosesolution could probably be attemptedin this context. In
the interior ofPannoniaa few late Romanfortified settlementsareknown, which
are strikingly uniform in size, methodof constructionand ground plan.37 The
best investigatedand best known is the fortification of Fenekpuszta 38 on the
302
south-westerncorner of Lake Pelso(Fig. 50), not far from the place where the
Sopianae-Savaria road, which playedan important part in late Romantimes as
part of the road from Sirmium to Trier, crossedthe river Sala. The walls of the
fortification enclosea rectanglemeasuring392 by 348 m, which is not compactly
built up and shows no indication that the settlementhad a military character.
Inside the fortification a large horreum,a Christianbasilicawith many periodsof
construction (PI. 38a), and a number of larger and smaller buildings were
exposed,amongthem a housewhich in size and groundplan can betakenfor a
palatial seat.The defensivewalls were very thick (2·60 m, with an evenbroader
foundation); the corner-, gate-and side-towers,altogetherabout forty-four in
number,were of large completelyexternalcircular plan, and the two gates,on
the north and south sides respectively,were also fortified with an inner court-
yard. To the fortification belongedextensiveand comparativelyrich cemeteries,
in which a Christiantomb-chapelwas also excavated. 39
The lay-out of Fenekpuszta,whoseancientnamewas probablyValcum, hasa

good parallel in the fortification of Gamzigrad(Fig. 50) in Dacia Ripensis,the
best-preservedand most imposing Roman monumentalbuilding in our pro-
vinces (PI. 37).40 Gamzigrad,too, is a rectangle(c. 300 by 230 m) enclosedby
mighty defensivewalls, which was only loosely built up. Up to the present,only
one building has been uncoveredinside Gamzigrad,a large palaceappointed
with all imaginableluxury: mosaicsfrom the sameschool as the late Roman
mosaics of Pannonia,rich architecturalplastic art, sculpturesof marble and
porphyry, a large number of expensiveobjects of use and so on. Specially
noteworthyis the themeof the mosaicsin the largestroom in the palace:this
mosaicfloor is divided into hunting-scenes,with the hunterswearing a shield-
shapedbadge on the upper arm like the hunters on the hunt-mosaicat the
imperial villa of PiazzaArmerina. The immenselystrong fortification of this
seatalso suggeststhat its owner was the emperor.The defenceswere similar to
those of Fenekpuszta,very thick walls (3-4 m wide), huge round towers
(twenty in all), two gatesand so on. Establishedas it is not far from a famous
thermalbathingestablishment,this fortification canperhapsbe interpretedas an
imperial hunting-castle,built in a fairly remote but pleasantregion for the
leisure of the emperors.Thoughit falls outsidethe scopeof this work, it must
be mentionedthat Gamzigradwas rebuilt under Justinianand chosenfor the
bishop'sseeandadministrativecentreof the xwpa (chora) of Aquae. The name
Aquae is evidently connectedwith the thermal springs of Gamzigrad.
Gamzigradwas also presumablythe centre of an imperial estate,probably
consistingmainly of forests and hunting-grounds.Its relationshipwith Fenek-
pusztais striking and must surely be consideredin any interpretationof this
30 3
Kornye
Fe.ne.kpuszta
fJ
fortified fortified
fortified
fortified
0 metres 300
I I fortified
fortified
fortified
fortified
Palace
fortified fortified fortified

fortified
Figure fO Late Romanfortified settlements

Pannonianfortification. The greatpalaceof Fenekpusztahas unfortunatelynot
yet beenexploredin an expertmanner;the finds so far madein the whole of the
settlementare however poorer than those of Gamzigrad,and do not suggest
imperial luxury. What do qualify the fortification of Fenekpusztafor the cate-
gory of imperial possessionare, first, the size, which is excessivefor anythingof
a private or military kind, and, second, the buildings inside the rectangle
enclosedby the walls, which suit neithera largecivilian settlementnor a military
camp.Fenekpusztawas not a town, nor was it a fort for somebranchof the late
Romanarmy. The largehorreum,which in itself might be imaginedas belonging
to a largelate Romancity or to a largegarrison,must accordinglybe interpreted
as indicating that Fenekpusztawas the centreof a latiJundium. The fact that this
latifundium cannothavebeena private one is shownby its not being uniquebut
representinga distinct type of late Romanfortification in Pannonia.
Apart from Fenekpuszta,large late rectangularfortifications with thick walls
rangingfrom 2.30 to 2·60 m in breadthandwith many roundtowersare known
from four further places,all in the interior of the province and lying on the
diagonal roads that connectedsouth-eastand north-west Pannonia. On the
Sopianae-Brigetioroad, about half-way betweenSopianaeand Lake Pelso,lies
the fortification of Hetenypusztanear Golle (ancient name, Jovia);41 and on
the sameroad southof Lake Pelsois found the fortification whoseancientname
is very likely to have been Tricciana.42 This fortification (modern Sigvir)
measures292 by 268 m; so far only its cemeteryhas beeninvestigated,and this
is one of the richestand largestlate Romaninhumationcemeteriesof Pannonia.
Herealso Christian grave-chambershave beendiscovered.43Approximately on
theline of thesameroadlies the fortification of Kornye (ancientnameunknown);
although this also has thick walls (2.30 m) and large round towers, it was
probably of smaller size than the others,perhapsc. 150 by 200 m.44 Lastly the
fortification ofMursellamust be mentioned,which lies on the Savaria-Arrabona
road; here too, only the cemeteryhas so far beeninvestigated.45 A noteworthy
characteristicof all thesefortifications is that they seemto have beencarefully
designed.Not only are the straight walls and round towers (which are always
tangentialto the wall) identical (Fig. 50), but so are the corners which form
exactright-angles.This latter characteristicrequiresespecialemphasis,for town-
and fort-walls elsewhererarely and indeed exceptionally show walls at exact
right-anglesto eachother.
We may justifiably doubt whether such uniform planning in the lay-out of
fortifications could have beencarriedthrough on private estates.Walls of such
extraordinarywidth, exceptionalevenin late Romanforts and fortifications, are
not likely to havebeenusedin private buildings, for the reasonthat fortification
30 5
of private villas towardsthe end of the fourth centurywas still only exception-
ally permitted. And, as we have pointed out, they cannot be consideredas
military constructionsor as newly built urbansettlementseither.It is all the more
noteworthythat we have datedimperial edictsfrom Mursella and Tricciana;46
and this allows the inferencethat when the occasionaroseemperorsdid reside
within thesefortifications. And since the edict of Constansfrom Mursella is
datedto 339, we may date their constructioneither to the Tetrarchyor better
still to the reign of Constantinethe Great. The coins found seemto start under
Constantine.
At Gamzigrad,the fort-like constructionof an imperial hunting-castlecan be
fairly accountedfor becauselate Romanimperial seatswere often regardedand
furnished as castra. In this respectDiocletian set an examplewhen he had his
palaceat Salona(Split) built as a luxurious fort. The defensivecharacterof sites
such as Fenekpuszta,which do not seem to have been built specifically as
imperial residences,is less understandable.To point out the increasingdanger
from barbariansin the borderprovincesof Pannoniais not convincing,particu-
larly if thesefortifications had really been built in the first half of the fourth
century. It is, of course,by no meansimpossiblethat large centresof imperial
latifundia alreadyexistedunderDiocletian or Constantineand were given defen-
sive walls later on whenfacedwith the increasinglydangerousforeign situation
on the Danube.The possibility also existsthat the estatesof which thesefortifi-
cationsformed the centreswere intendedfor the settlementof barbariansfrom
the left bank of the Danube,who were transferredunder Diocletian or Con-
stantineto Pannonia,and that the centresof the estateswere heavily fortified in
order to intimidate the barbarianswho had been acceptedinto the empire. It
might be relevant that Fenekpusztaand Sagvar weresituatedat Lake Pelso,
which was regulatedby Galerius just after 295 when the Carpi had been give
land in Pannonia.A settlementof Carpi is moreoverattestedaroundSopianae,
in the generalareaof which Hetenypusztalies. It is hopedthat the excavations
started at Fenekpuszta,Hetenypusztaand Sagvar will be able to provide an
answerto thesequestions.
It is quite possiblethat further agriculturalcentresexisted,now unknown,of
the type of Fenekpuszta:apartfrom the latter, which hasbeenwell known for a
century,fortifications of this type have cometo our knowledgeonly during the
last few decades.So we cannotexclude the possibility that the imperial estate
with a fortified centreconstitutesan estate-typein Pannonia.Even the sparse
epigraphicmaterialfrom late Romantimes doesyield one pieceof evidencefor
imperial estates:this is the silvae dominicaein the surroundingsof Savaria47-but
an administratorof the resprivata is attestedonly in the provinceof Savia.48 A
306
numberof place-namesof varying age can be attributedto imperial estatestoo,
for instanceCaesariana,which is to be soughtin the surroundingsof the large
villa of Balaca,and a number of places with the names Jovia or Herculia, of
which howeveronly somecan be localized (e.g. Hetenypuszta).
Fortifications of Fenekpusztatype have a further important relevancefor
Pannonianhistory. Behind their strong walls in the early Middle Ages there
survivedremainsof the Romanpopulation,and otherethnicelementswhich are
not easy to define. These were sometimesresponsiblefor an archaeological
'culture' of specialtype, suchas the so-calledKeszthelycultureof the inhabitants
of Fenekpuszta. 49 To all appearance the fortifications wereplacesof refugemuch
in demandduring the Dark Age, and were thereforeoften capturedand chosen
as residencesby successivemastersof the country. We shall later return to this
aspectof them.
As hasbeenseen,the developmentof largeestatespartly in private but mainly
in imperial possession was a significanttrendin the Danubeprovinces.Up to the
middle of the third centurylarge estatesplayeda certainrole only in the interior
of Pannoniaand UpperMoesia,in the wider hinterlandof the limes. But evenin
someterritories of the hinterland,as for instanceat Scupi, municipal medium-
sized estateswere preserveddown to the time of the Severi,50and the special
circumstancesof the mines of Upper Moesiaalso resultedin tendenciesthat are
different from 'normal' development.For these reasonsthe establishmentof
large estatesover most of Upper Moesiacan be regardedas a processcausedby
late Roman development.
If our hypothesisabout the conditions necessaryfor the preservationof the
estatesof small peasantsin the border area is correct, that is if local military
recruitment in the Severanperiod did indeed maintain the prosperity of the
small estates,then it must be concludedthat the soldiers' living conditions
underwentfar-reachingchangestowardsthe end of the third century or a little
later, and that thesechangescontributedtheir shareto the formation of large
estates,or were even responsiblefor it. It is beyondthe scopeof this work to
outline the military reforms from Gallienus to Constantine; their essential
characteris presumablywell known even if their details and in particular the
stagesof their developmentare still obscure.Thesereforms were doubtlessas
effective in Pannoniaas elsewhere;as a result the frontier troops lost their old
political role and also suffered material handicaps.It was the newly created
mobile troops in the larger centresin the interior of the empire who received
political and materialprivileges: theseplaceswere also centresof administration
and so no distinction can be madebetweenmilitary and civilian prosperity.
In the fourth centurythe centresof prosperityand wealth were no longer the
30 7
military areasalong the Danube,but the agricultural estatesin the countryside
andthe centresof administration,which also lay in the interior of the province.
The towns in the fourth century

The developmentof large estateswas presumablythe most important cause,
or at leastone of the most importantcauses,of the declineof municipal life. Till
the middle of the third centurythe curial upperclassof the townshadremaineda
socialgroupprosperous,active,preparedto makesacrifices,and-inthe frontier-
zone-alsoclosely interlinked with the military: this classis hardly heardof in
the fourth century. Under the Tetrarchythe decurionsor higher magistratesof
the towns hadindeederecteda few dedicationsfor the welfare of the emperors51
in conformity with duty, but after this they disappearfrom the inscriptions; so
do the soldiers,who had beenthe other social classof the provinceto be active
epigraphically.The loss of their old role by the curial classandfinally by munici-
pal administrationsis attestedby an inscriptionfrom Savaria,which recordsthat
the horrea of the town werereplenishedunderConstansandthe supplysecured.52
On this inscription, dealingwith a purely municipal affair, neitherthe municipal
magistratesnor the colonia Claudia Savaria are named. On the late Roman
inscriptions of the province only officers of the imperial administrationand
dignitaries of the new aristocracyare named: soldiers occur but rarely, and
municipal magistratesnot at all. 53 Possibly this is to be accountedfor by the
customof Christiansto omit mentionof the worldly position of themselvesand
their relatives from grave-inscriptions(Figs 55, 56), but it is more likely that
holdersof the traditional offices no longer erectedinscriptions:they were, after
all, the socialclassmost tenaciousof traditions-andof the old religion amongst
others. Those who set up inscriptions in the late Roman period were, as has
beensaid, membersof the new aristocracy,and the distribution of inscriptions
suggestsa considerablesocial displacement,for late inscriptions mainly come
from the interior of the province (Fig. 5I), not from the border area. The
inscriptions themselves, of course, do not allow any further conclusions,
especiallyin view of their generalrarity. The importantpoints are whereand by
whom inscriptionswere still erected;the answersto thesequestionspoint to a
social displacement. 54
The curial upperclassdid not of courseceaseto exist. At the town of Emona,
which at that time had long belongedto Italy, the decurionswent out to greet
Theodosiusin the traditional toga,55 and even towards the end of the fourth
century the administrationof the mines of Upper Moesia was entrustedto the
chargeof the curiales, or, more correctly, to procuratorschosenfrom amongthe
308
fo
rm
ed
formed formed
form
ed
o 100 200 k
m
Figure J I Distribution of Christian inscriptions
number of the curiales,56 This office presumablyformed one of the numerous

burdensthat the curialeshad to bearin the late Romanperiod, It is evenpossible
that someself-governingtowns were not founded before the beginningof the
late Romanperiod, for instanceJovia, formerly namedBotivo, eastof Poetovio
on the Drave, and possibly some of the late Pannonianand Upper Moesian
30 9
towns such as Remesiana,the seeof a bishop in the fourth century; Sopianae,
seatof the praesesof Valeria; Aurelianumor Aureliana,etc. The prerequisitefor
founding a town was still the existenceof a prosperoussocial class capableof
becomingthe class of curiales.
It was presumablyto these curial estate-ownersthat the smaller but often
quite luxuriously appointedvillas in the surroundingsof Aquincum belonged,
which were of fourth-centurydate.57The neighbourhoodof Aquincum is quite
well investigated,which enablesus to perceivea numberof other symptomsas
well. The emigrationby the rich inhabitantsof the town into the surrounding
territory could be nicely proved. The late antique cemeteries of Aquincum are
very poor, but thereare rich individual burials, includingmummifiedinterments
and sarcophagi,in the wider circumferenceof the town.58 It could also be
observedthat building activities at Aquincum continued only till about the
middle of the fourth century; after that, not only were no new buildings con-
structedbut the older buildings were either not kept in repair or they continued
in usein a very poor way. The hypocaustswere no longer employed,the rooms
were heatedby primitive hearthsand ovens,the drains that had collapsedwere
no longer repairedand so on.59 The housesand the gravesin the surroundings,
on theotherhand,attesta certainwealth,thoughthis wasno longer concentrated
in the town but in the countrysideinstead.
The resultsof researchin Aquincum canprobablybe generalizedfor the other
towns of the borderarea.Carnuntumis describedby Ammianusas desolateand
ruined, and excavationsthere have not so far disprovedhim.60 However, his
descriptionof circumstancesin Pannoniaat the time of Valentinian'svisit in 375
has given rise to controversy,since to begin with it is not certain whetherthe
symptomsof decayresultedfrom a longish processof declineor from the taxa-
tion policy of Probus the praetorian prefect. Some details of Ammianus'
description61 suggestthat in Savariaand evenmore in Sirmium everythingwas
still in goodorder,eventhoughthe town-wall was in a stateof neglectas a result
of the long period of security. When the Sarmatiansbroke in during 374, the
walls andditch of Sirmium hadto be repairedin greathaste-theditch was filled
with stonesthat had fallen from the walls-andfor this work money was used
from the municipal treasurywhich had beencollectedfor building a theatre.As
far as Savariais concernedwe learn that the gateswerein a tumbledownstateat
the time of Valentinian'svisit. In my opinion, we cannotconsiderthis the result
of long peace, becausenormal urban life cannot be imagined without busy
traffic through the gates.There is much to supportthe view that at Savariaat
least thesymptomsof decayare traceableto the taxationpolicy of Valentinian,
or rather to its cruel implementationby Probus.From Ammianus himself it is
310
clear that at that time the interior towns of Pannoniawere affectedby a tempor-
ary crisis whereasthat of the Danubetowns was chronic.
As already shown, the rich inhabitants of Aquincum-presumablythe
curiales-movedout of the town and had country-housesbuilt for themselvesin
the surroundingterritory, wherethey also had themselvesburied.Thetowns on
the Danube,formerly centresof prosperityand focal points of politics, became
impoverishedand presumablydepopulatedtoo, first becausethe richest and
most activeleaderswithdrewfrom urbanlife, and secondbecausethe soldiersin
the border zoneno longerconsistedof troopsof the first quality. The processof
concentrationof agriculturalestatesaccordinglyextendedover the whole of the
areaof the province. Neverthelessthere were towns able not only to preserve
their role but to experiencea new period of prosperity. This is all the more
curioussincethe towns in the interior of Pannoniahadalreadylost much of their
original importance,mainly as the result of the concentrationof estates.Now
thesevery towns becameflourishing centresonceagain.Not all of them,though;
if we take into accountlate Roman buildings known so far, the wealth of
burials andthe occurrenceof inscriptions,we can suggest that prosperityis only
characteristicof towns that for somereasonwere privileged by the government
(amongthem towns in the vicinity of Sirmium, and a few specialcasessuch as
Constantine'sbirthplace,Naissus),or else that were centres of the provincial
administration,suchas Sopianae,Savaria,Sisciaor Viminacium. It will be better
to speakof a further displacementof the centresof influencethan of a gradual
decline of the towns; this displacementcausedthe decayof someregions, but
not of the whole area.
This shifting of the focal points probably did not happenimmediatelyunder
the Tetrarchy. We have a few indications that life started to move again in
the old ways in the time when consolidationwas beginning. A few grave-
inscriptions,which are more crudely executedand poorer than tombstonesof
beforethe middle of the third century,must presumablybe datedto the time of
the Tetrarchy.62And we havea few altarsand building-inscriptionsdating at the
latest from the beginning of the fourth century that still retain the traditional
form and formulation, and the majority of thesecomefrom centreswhich had
playeda leadingrole in Severantimes.63 In the areaof the new Dacia, to which
the legions of old Dadahadbeentransferred,andwherethe Romanizedcivilian
population evacuatedfrom Dacia had been settled, there exist sculpturesand
inscriptionslater than the time of Aurelian; someof the ornamentaldetails and
over-all designof the stonesculpturefrom Dacia Ripensishave suggestedthat
evacuatedDacian stonemasonscontinuedto work in the new Dacia.64-
Underthe Tetrarchysometowns will havebeenprovided withtown-walls. So
3I I
far this is certain only for Scarbantia,where the wall-which remainedin use
virtually without alteration during the Middle Ages and is preservedto the
presentday (PI. 2ob)-couldbe datedin fairly recent excavations.65 The town-
wall of Scarbantiaencloseda roughly oval polygon. An areaof similar shapewas
surroundedby the town-walls of Bassianatoo; it is only known from air-
photography.66Certainlylate Romanare the town-walls ofUlpianum,Horreum
Margi andRemesiana67 andsomeothershardlyyet investigated68 thoughknown
to exist. However,when we rememberthat from Diocletian to Valentinian our
provinceswere fairly peacefuland that under Valentinianthe walls of Sirmium
werein a neglectedstateandthoseof Savariatumbling down, it may not be rash
to conclude that the walls were built under the Tetrarchy. But the proof is
availableonly at Scarbantia.
It is only under Constantinethat a radical changetook place.After Constan-
tine all archaeological,epigraphicand other evidencecombinesto show urban
prosperity in the centres of administration. Among these the bishops' sees
presumablyhave to be included; but most of theselay in the samecentresas
thoseoccupiedby the bureaucracy.At any rate, wealth is often expressedin an
abundanceof Christian monuments and in particular by richly furnished
Christian cemeteriesof which the date must be Constantinianor later.
The imperial city of Sirmium was given most of its monumentalbuildings
under the Tetrarchy(Fig. 29, p. 163). This conclusionis basednot only on the
disturbed conditions precedingthe accessionof Diocletian which would not
allow of large-scalebuilding, but aboveall on the choiceof Sirmium as a perman-
ent imperial residenceby Galerius.Licinius orderedbathsto be built there, for
which the column-capitalswere fetchedfrom a quarry on the Dalmatianisland
of Brae.69 WhetherthesethermaeLicilliallae can be identified with the truly huge
remains of baths (PI. 39a) uncoveredby the new excavations70 is as yet un-
decided.Thesebathsat Sirmium were constructedin building-tiles, as were the
other large buildings recentlydiscoveredat SremskaMitrovica;7I amongthem
werea largehorreum,part of a palace-likebuilding (PI. 39b), a hippodromeprob-
ably connectedwith the Imperial Palaceand, in the south-westcorner of the
town, a row of shopswhich possibly belongsto a market-hall. The remaining
excavatedbuildings are dwelling-houses,amongthem one of a distinctly upper-
classcharacter.Buildings constructedof tiles are very rare in Pannonia,and this
is all the more striking sincein many areasno building-stoneis available.There
are large districts wherestonesfor building had to be fetchedfrom quarries50
km away from the building-site.Presumablythe reasonwhy, exceptionally,tiles
were usedin Sirmium was that the very pretentioustechnicalexecutionof the
public buildings was only possiblein tile. Of greatsignificanceare the Christian
312
cemeteriesof Sirmium (Fig. 52). Thesewere excavatedin the nineteenthcen-
tury, unfortunatelywith little care (but perhapswith too much imagination).
They have yielded a large number of grave-inscriptionsand an extraordinary
variety of subterranean buildings.72 Below we shall return to this cemetery:here
it need only be mentionedthat similar cemeteriesare equally characteristicof
the other administrativecentres.
Naissusexperiencedits first real prosperity only in the fourth century, not
becauseit was an administrativecentre but becauseit formed an important
nodal point in west-eastcommunicationsand was also the birthplaceof Con-
stantinethe Great. StephanosByzantioswas probablycorrectly informed when
he consideredthis town to be the 'foundationand the home of Constantine'.
Although Naissushad becomea town as early as the reign of Marcus its real
founder was evidently its greatestson, to whom a bronzestatuewas set up not
far from the bridge over the Nisava.73 Of the cemeterieswhich extend over a
very considerablearea east of the town some parts have been laid open in
several stages.74 The large cemeteriesof the late Roman town are very uni-
form. Here,too, subterranean tombswhosewalls arebuilt of tile or stone,grave-
chapelsaboveground,evenfunerarybasilicasand grave-chambers with frescoes
are frequent and prove not only the prosperityof the town but its size also.
There exists a very close relationshipbetweentwo painted grave-chambers
from Naissus75 and Sopianae,76which was also the seatof a provincial admini-
stration.From Sopianaecomethe most importantearly Christianmonumentsso
far discoveredin Pannonia(Fig. 57, PI. 40). The grave-field (Fig. 53) is as rich
and shows as much variation as the cemeteriesof Sirmium and Naissus.77
Finally we must mention the cemeteryof Savaria,78the capital of Pannonia
Prima: it hasnot so far beeninvestigatedat all but is unequivocallyattested,and
to judge by its inscriptions was probably as rich as the others.In Savaria,too,
the remainsof a large building were uncoveredthirty years ago, containing a
large hall with an apse;its floor had beenadornedwith first-class mosaics.The
walls were coveredwith marble slabs,and the whole hall was provided with a
hypocaustedheating system. The mosaicscan be attributed to the workshop
which at the beginning of the fourth century produceda number of mosaic
floors for the palacesand Christian cult-buildings of Illyricum and Aquileia.
The excavationof this building in Savariawas from the start carried out on
the assumptionthat it was the Christian basilica, built under Constantine
for the martyr-bishopQuirinus. The doubtsvoiced during the last decadeabout
this identification haveonly recentlybeensystematicallyarguedand at the same
time the hypothesishas been suggestedthat the hall with its mosaicswas the
aula of the imperial palacein Savaria.This palacelay in the immediatevicinity of
SIRMIVM SJRMIVM
SynerotasCemetery East Cemetery o .30
metres
VLCIS'IA CASTRA AQVINCYM
Figure 52 Plansof Christian cemeteries
the forum of the colonia,as is shownby fragmentsof the statuesof the Capitol-
ine Triad which were found in the southernneighbourhoodof the aula.79
Apart from Sirmium, Sopianaeand Savariawhich were seatsof the provincial
administrations,and apart from Naissus,the birthplace of Constantine,other
towns in the interior of Pannoniaand Moesiadisplay the uniform characteristics
of late Romanmunicipal prosperity.The recently exploredChristiancemeteries
of Ulpianum (which was alreadycalled Ulpiana)-thequartersof a legio pseudo-
comitatensisaccording to the Notitia 8°-sporadicburials around Scupi which
31 4
often employ older tombstonesas packing,81 the grave-chambersin the sur-
roundingsof Remesiana 82 and the gravesofViminacium83 all haveexactparal-
lels in Pannonia.There are no differencesin the rite or the combination of
grave-goods,but only in the wealth of the burials, and in this respect,too, the
differencelies not betweenPannoniaandUpperMoesiabut betweenthe interior
ofPannoniaand of Upper Moesiaon the one handand the frontier zoneof both
provinceson the other. In the towns of the frontier zonethe burials are poorer;
the gravesare mostly simple earth-gravesor graveswith tiles, burial-vaultsare
rarer,andstructuresornamentedwith paintingsor mosaicsevenmore so. This is
equally characteristicof the towns and of the settlementsat military sites on the
frontier; whilst in the hinterland, as well as rich cemeteriesat the towns, a
numberof wealthy single gravesare also known from non-urbansettlements. 84
Through the archaeologicalindications of prosperity in the towns in the
interior of the province a further important difference can be traced. We have
already pointed out that the majority of towns of late foundation were not
towns in the propersenseof that word, for the social, economic-andoften the
geographicalconditions as well-for implanting regular urban life were not
available.And these'towns' were the onesto display no tracesof prosperityin
the late period, even though they may have beensituatedin the interior of the
province. The reasonwas not simply that for the seatsof provincial administra-
tion, and presumablyfor camps of the field army, traffic-centresand thus the
sites of genuinetowns were chosen.In this respectit is characteristicthat all
three pseudocomitatensian legions in Upper Moesia were placed in important
places of the country: Timacenses,Ulpianensesand Scupenses. 85 And the
developmentof large estatesmust also be consideredan important factor. It is
possiblynot without importancethat one of the greatfortified latifundia centres,
discussedabove,was built in a municipium of Hadrian. The distribution of late
Roman-earlyChristianinscriptions(Fig. 5I) is limited either to towns (Savaria,
Siscia, Mursa, Cibalae, Sirmium, Viminacium, Ratiaria, Naissus) or to areas
where there had beenno regular towns even in earlier times: at the latter these
inscriptionscometo light in unimportantsmall places,evidentlyburial-placesof
owners of large estates,or in the estate-centres.
As we havealreadystated,late Romanburial-ritesare strikingly uniform. The
sametendencytowardsuniformity canperhapsalso be discernedin otheraspects
of daily life. One characteristicis particularly striking and can be further
establishedby excavation: as examples we may take the settlement at the
Pannonianlimes-fort of Matrica (Szazhalombatta)and the large settlementof
Gorsium(Tac) in the hinterlandof Matrica.86 In both settlementsa successionof
layers extendingfrom the secondto the fourth century came to light. In the
31 5
Thefinal period oj prosperi~y
/cf==~~_O /cf==~~_O
FenekpU51.ta metres
AQVINCVM O.,..__/cf==~~_O /cf==~~_O.._3.f )(
/cf==~~_O
SOPIANAE f...__T.t:,O==2::1..o_ _3..~ metres
Figure J3 Christian basilicae: funerary monumentsin Sopianae

secondcenturyboth settlementsconsistedmainly of dwelling-housesandbuild-
ings not constructedfrom stoneor tile. Stonebuildings appearfirst in the third
centurybut becomegeneralonly in the fourth, so much so that onecould reckon
that all dwelling-houseswere stone-builtat that period (though the worsening
quality of the stonework cannot be denied). At Matrica (Fig. 54) remains of
severalsmall houseswere discoveredwhosewalls had no foundations,nor were
the stonessetin mortar. Dry-stonewalls without foundationdisappearonly too
easily, and so they cannotalways be tracedin excavations.This is presumably
the reasonwhy late Romanvillage settlementsor, in general,simple dwelling-
housesare only seldomtraceablearchaeologically.The dwelling-housesof the
earlier periods were often dug partly into the ground and settlementsof early
imperial date are marked by the great numberof pits. Settlementlayers of the
fourth century haveleft no such obvious tracesbehind. The numberof stone-
built housesof fourth-centurydate establishedby minor excavations,however,
is not negligible,87and we can thereforeassumea generalchangein building-
methodin the fourth century.
The barracksin late forts also probablyhad dry-stonewalls without founda-
tions. Various reconstructionsand additionsin the left half of the praetenturain
the legionaryfortressat Carnuntum88 and the interior buildings of the recently
investigatedlate Romanfort of Tokod (PI. 36a),89consistedof dry-stonewalls.
In urbanliving-quarters,for instanceat Aquincum, dry-stonewalls also become
frequent.Bonding coursesof tiles can also be considereda technicalinnovation
in construction;theseare characteristicof late Roman military buildings and
presumablycivilian architecturederived them from that source. The courses,
two to four tiles thick, were built into the walls at intervals of about I m, in
order to strengthenthe rubble masonry.
This change in the way and standardof life presumablyfollows certain
changesin the social structure.But investigationsso far undertakenare insuffi-
cient for full comprehensionof these,nor are our sourcessuitedfor the answer-
ing of suchquestions.As the epigraphicsources,which yieldedrich information
up to the middle of the third century,dry up, greaterreliancehasto be placedon
purely archaeologicalresults,and this we have attemptedto do. We have been
ableto show the probability that the municipal upperclasswhich had suchclose
connectionswith the army was replacedby an aristocracyof owners of large
estates;these men may have been partly derived from the former municipal
upperclass,and in the fourth centurythey did include curiales in their number;
but in generalthey were much more closely linked with membersof the new
bureaucracythan with the military on the limes. And among owners of large
estateswe may assumethat the emperorhimself had by far the largestposses-
31 7
Thefinal period oj prosperity
2nd century
3rd century
4th- century
o
.' 10 20 30
~' meues
Figure 54 Plan of the vicus of the auxiliary fort at Matrica

sions in land. But a social structureof that kind would necessarilyresult in a
social polarizationsuchas had hitherto beenunknownin the provinces;yet to
judgefrom the standardizationof the way of life which we havedescribed,this
polarizationdoesnot seemto have occurred.
Economicand social developmentsin the late third andfourth centuries

Neverthelessit may be suggestedthat there did exist a sharperpolarizationin
our provinces than in former periods. In the first place it is significant that
hoardscomposedof objects of preciousmetal and of works of art are mainly
fourth-century or were buried at that time.90 The rich graves of the fourth
century, that is thosefurnished with preciousgrave-goods,are wealthier than
graves consideredrich in the earlier periods. The luxury of the wealthiest
dwelling-housesincreases,yet at the sametime theseluxury housesare met with
more rarely. Furthermorea gradual processof polarization can be traced by
analysisof late Romancemeteries;fourth-centurygravesexist in a quantitysuffi-
cientfor statisticalinvestigation.It hasbeendiscoveredonly recentlythata gradual
pauperizationcan be tracedin late Romancemeteries,not in the sensethat all
gravesbecameincreasinglypoor, but becausepoorly furnishedgravesandthose
without grave-goodsbecamemore and more frequent,whilst the wealth of the
few rich gravesremainsthe same. There are, however, marked differencesin
wealth betweenindividual cemeteries;in general,cemeteriesin the interior of
Pannoniaare richer thanthoseof the settlementson the limes, while outstanding
wealthis seenin the cemeteriesof administrativecentresand thosebelongingto
the centresof fortified imperiallatifundia of the type of Fenekpuszta. 91
The standardizationof the way of life is thus not basedon a levelling of

classes.It showsitself in characteristicsindicating increasedindustrial produc-
tion ratherthan actualwealth. In Chapter7 it was pointedout that local charac-
teristicsin the pottery and bronzeindustries,in burial-ritesand in costumeand
so on, suddenly faded out in the Severanperiod. Provincial culture became
colourlessand undifferentially 'Roman',and this is clearly seenin different fields
and manifestationsof culture, from pottery to the art of the stonemason.This
processcontinuedin the fourth century.It hasalso beenshownin studiesoflate
Roman cemeteriesthat with very few exceptionsthe grave-goodswere pro-
ducedon the spot; this could be provedonly by mappingthe minor peculiari-
ties. This may be illustrated by a few examples.Bone braceletsare amongthe
mostfrequentgrave-goodsin virtually all cemeteries.At first glancethey are all
very similar. But the simple decoration-incisedcircles and dots-showslocal
characteristics.Jugsand beakersof pottery are also frequent grave-goods.The
31 9
jugs mostly havea handleand a bulbousbody, and the beakersare of types that
are closelyinterrelated.Thereare howeverhardly any typesthat havea distribu-
tion over the whole of the province.A fine typology demonstrates that eachtype
is characteristicof a limited areaonly. Moreover,in cemeteriesround which no
suitableclay was available,pottery is rare, but in othersthe majority of grave-
goods are ceramic. When classified closely, differentiation of grave-goods
increasesduring the fourth century, though if an insensitivetypology is em-
ployedthe materiallooks very uniform. The productionof local workshopswas
everywhereable to satisfy demandfor cheapobjectsof use;the waresproduced
were very simple and mass-produced,attesting uniform poverty among the
population.92
The wares not producedon the spot are very few in number. Chief among
them must be consideredthe cross-bowbrooches(Zwiebelknopffibeln) which
are also very uniform and which were probablyproducedin centralworkshops.
Thereare someindicationsthat this type of broochwas reservedas a badgefor
certain ranks in the administration. From central workshops, too, came the
bronzemountsdecoratedwith figures, usedon small caskets(PIs 44b, c); some
of theserepresentations refer directly to the emperor,andpossiblyhad a certain
official character.Of the remainingrepresentations someare decidedlyChristian
in character,the remaindertakenfrom paganclassicalsymbolism.93 The remain-
ing imports are mostly luxury goods, such as glass vesselswith cut ornament
probably from Rhenish factories,94 objects of jet also from the west,95 and
finger rings with intaglios,bucklesdecoratedin niello, andso on. The contribu-
tion of the eastin theseluxury goodswas presumablylarger than can be proved
by available parallels; inscriptions where they occur on such objects are
strikingly often in Greek.
Some of the objects which reached Pannoniafrom distant sources were
officially provided for the use of certain ranks and cannot be included among
evidencesof genuine trade. Import-trade was restricted to luxury items the
categoriesof which were probablymuchmore limited thanthosehandledby the
busy commerceof Flavian to Severantimes, when import in bulk is proved
aboveall by samianware (terra sigillata).
What has beensaid of courseappliesto Pannoniaonly, sinceno studiesexist
of excavatedlate Romancemeteriesin Upper Moesia,which are in any casefew
in number; nor do we possessmassivepublications of finds from excavated
settlements.It will, however, be permissible (with certain reservations)to
extendthe resultsobtainedin Pannoniato Upper Moesiatoo; this is becausethe
cemeteriesthat so far have come to our knowledgedo not differ, in rite and in
the grave-goods,from those in Pannonia.A uniformity covering both pro-
320
vinces can thereforebe consideredlikely. This was, however, a uniformity of
broadsectionsof the populationpossessingin the maina ratherpoor standardof
living. The small upper classwas richer than the larger municipal and military
upperclassof former times, but, perhapspreciselybecauseof its small numbers,
it was the bearerand representativeof the uniform upper-classculture of the
empire which now rarely yielded any local characteristics.
Becausethe finds do not admit an import in bulk, we are facedwith the prob-
lem how to explain the most striking peculiarity of late Romanmonetarycircu-
lation in Pannoniaand Upper Moesia. It has long beenknown that coins from
Balkan and easternmints form a very considerablepart of all the hoardsand
isolatedfinds, whereasItalian and westernmints are hardly representedat all;
not evenAquileia hasa percentageworth mentioning.96 Perhapsit was the high
easternproportionin the coinageof Pannoniathat inducedhistoriansto suggest
that late RomanPannoniaenjoyedan increasingimport-tradefrom the east.But
the presentstate of knowledge offers no causefor that. The east had always
played its part in imports to Pannonia,and in the fourth century, as already
mentioned(p. 320),therearestill anumberof articleswhich mustbe derivedfrom
the easternprovinces.For example,thefew finds of textiles from fourth-century
graves,with silk and linen representedamongstthem, presumablycamefrom
the east.97 But there is no archaeologicalevidencethat the east achievedpre-
dominancein the trade of Pannonia,and we must thereforeassumeeither that
oriental imports were composedof organic material and are therefore un-
traceablearchaeologically,or that Pannoniawas paid in coin for exportsto the
east;for this secondsuppositionthe coin-finds can be cited in evidence.
Here we would have let the question drop were it not that a remarkable
sourceoffers groundsfor further hypotheses.This sourceis the Expositio totius
mundi, a mainly commercially orientated survey of the then known world,
written c. 350 by a Syrian. The authorknows hardly anythingof the tradeof the
westernprovinces;evenprovincessuchas Gaul are dealt with in insipid general
remarksand political details. It is all the more striking that he considersPan-
nonia to be 'a countryin all respectsrich', having a busytradein corn, cattleand
'in part' in slavestoo. We must attachparticularimportanceto this information
becausethe countrieson the lower Danube(MoesiaandDaciaRipensis)though
famed in the empire for their agriculture,accordingto the Syrian author were
only self-supplying;that is, they did not export.Whetherthe renewedattacksby
the Goths from the middle of the third century onwards and the subsequent
Gothic settlementswere responsiblefor the decreaseof agricultural exports, I
am not able to decide; it is howevernot unlikely that the reasonfor Pannonia
beingableto enterthe tradeof naturalproductswasthat theprovinceswhich had
321
a bettergeographicalposition,commerciallyspeaking,wereno longerableto ex-
port. The prerequisiteof productionfor exportwas presumablythe largeestates.
The Syrian author of the Expositio names Noricum as supplying garments
(Norica vestis) in the samechapterwith Pannonia;this might perhapssuggest
that the Norican export went via Pannonia,and this makes senseonly if it
travelled down the Danube towards the east. In the same way, Dardania,
exporting Dardanian cheeseand lard, is mentioned only in passing in the
chapteron Macedonia;the reasonbeing that it had beenMacedonia,in fact, that
had tradedtheseDardaniangoods.
Apart from corn and cattle, slaves are named as exports from Pannonia.
Where theseslaves came from we do not know; primarily we might imagine
that internal unrestin the Lowlands beyondthe Danube,the frequent struggles
betweenSarmatians,Gepidae,Vandals and Goths,98 might generatea slave-
trade on the frontier through the sale of prisonersby the Sarmatiansinto the
empire. Regular trade on the bank of the Danubeis attestednot only by late
Romansmall coinage,which is very frequent in the HungarianLowlands, but
also by the curious inscription from the time of Valentinian, reporting that a
burgus was erectedfor the purposeof border trade.99 In the fortified landing-
place on the left bank of the Danubenear Nogradverocepiecesof raw amber
were foundIOO attesting the continuanceof the amber-tradethrough the bar-
barian approachesof Pannoniain the fourth century.
Accordingly it is possible that Pannoniain the fourth century exported
agrarianproductsand slaves,and that the easternprovinceswere interestedin
this export. The only relevanttraffic-route was the Danube,on which the goods
weretransportedeitherto RaetiaandNoricum or to the Black Seaandultimately
to Asia Minor. Even southernPannoniahad accessto the trade of the empire
only via the Save and the Drave which ran eastwards,and so we need not
confine our searchfor the large estatesresponsiblefor the export to the vicinity
of the Danubein Valeria and PannoniaSecunda.
The fact that Dardania exported the products of its mountain farming-
cheeseand lard-to Macedoniawill not be surprising seeing that the Axius
was a convenientandat the sametime the only water-routeconnectingDardania
with the Mediterranean.
Religion in the late third andfourth centuries

The trade-relationswith the east were presumablypartly responsiblefor the
fact that oriental immigration did not ceaseafter the crisesof the third century.
On Christian grave-inscriptionsChristiansnewly arrived from the easternpro-
322
vincesarerepeatedlynamed(Fig. 55); orientalGreekparticipationin the earliest
Christian communitiesseemsto have been very great. The large majority of
Christiansknown by namehadGreeknames;IOIso did the bishops,the majority
of whom had Latin names only from the middle of the fourth century on-
wards.I02 But even among Christians with Latin nameswe may assumethere
were Greeks;oneof the earliestDanubianbishops,Victorianus,was a Greek.103
The earliestChristiancommunitieson the middle Danubewere small groupsof
people from the easternprovinces. They are found mainly in and around
Sirmium104 where Orientalsattractedto the headquartersof the Illyriciani and
later to the imperial court had gladly settled. The legendarylist of bishopsof
Sirmium goes back to the disciples of the apostles.The real circumstancesand
prerequisitesof the Christian mission in our provinces can presumablybe
recognizedin the attractionexertedby Sirmium, despitethe fact that until the
end of the Tetrarchy the policy of the Illyriciani was decidedly hostile to
Christianity.
Paganism,indeed, maintaineditself tenaciouslyin our provinces,and when
Constantinestayedin Sirmium in the pos and 320S he probably had to meet
religious oppositionfrom the Illyriciani of the Licinian party. After Constan-
tine's victory over Licinius the pagansin our provinceswere silenced,as is best
illustrated by the suddenceasingof pagan dedicationson inscriptions. Since
conservativepagan elements were especially distinctive of the circle of the
curiales, one of the reasonsfor the disappearance of epigraphicevidencefor the
curiales may havebeenthe disinclinationof pagancuriales to expresstheir beliefs
either on votive inscriptions or on tombstones.At any rate the drying-up of
inscriptionscannotbe simply accountedfor by a changein tasteor by economic
decline. As we have already stated,pagan dedications,and grave-inscriptions
too, are still found eventhoughnot in large numbersunderthe Tetrarchyafter
the consolidation;in their suddendisappearanceunder Constantine,we must
recognizepolitical and religious reasons.For at this sametime the Christians
start to set up gravestonesin certain towns;105 it is inconceivablethat none of
the curiales should have had the material meansof setting up a tombstonefor
himself. The distinctly luxurious burials however are indisputably Christian,
and gravemonumentsandgrave-inscriptionsareerectedonly by Christians;and
this provesnot only the gradualChristianizationof the wealthy upperclassbut
also the assuredand even challengingcharacterof the new religion. Since the
attitudeof the Illyriciani up to the time of their defeatwas hostileto Christianity,
we must assumethat the ruling classin our provinceswas paganup to the time
of Constantine,and attribute their conversionto his reign; no doubt political
considerationsplayed the main part in the change,as also in the choice of the
32 3
new aristocracy.Before we enteron the ecclesiasticalhistory of our provincesit
will therefore be worth while to set out the evidenceconcerningthe end of
paganism.
Under the Tetrarchy paganismwas still flourishing in Pannoniaand Upper
Moesia. Official votive inscriptionsset up by municipal magistratesare, in fact,
rather colourless-dedicationsto Juppiter Optimus Maximus are the most
frequent-butthereis good evidencethat sanctuariesof Mithras were still used
and maintained; indeed, the cult of Mithras may have enjoyed the greatest
popularity of all. This at any rate is suggestedby the action of the imperial
participants in the conferenceof Carnuntum in 308, who dedicatedto this
particulargod a largestone(thoughin fact re-workedfrom anold altar (PI. 34b)),
and who restoreda Mithraeumat Carnuntum. 106 To all appearance
the Mithraic
sanctuariesfell victim only to the destructiverage of the Christians;this can be
deducedfrom the fact that often all the trappingsof the cult had beencollected
andhiddenin the ground,107showingthat they were still in useat that time. It is
well known that somepaganrites and beliefs managedto survive still later, and
it is significant that a small relief of the Danubianpair of Rider-godswas found
in a layer of Constantiniandate,and a similar small relief even cameto light in
the imperial villa at Mediana.l08
Finds of courseexist that prove the use of pagancult-placestill late in the
fourth century. Least significant are the coins thrown into springsand wells,109
which often go right down to the end of the Romanperiod. Magical texts and
similar superstitiousfinds do after all turn up evenin Christian contexts,IIO and
this sort of superstitionis behind the offering of coins to water. But as late as
the reign of Gratiana Mithrael..lm seemsto havebeenthe sceneof religious rites
not far from Scarbantia. III As far as can be established,pagan sanctuariesall
suffered violent destruction, from the capitolium of Scarbantiato individual

altars standingin the open,which were often buried in the groundII2 or walled-
up.II3 The smashedpieces of the colossal statuesof the Capitoline Triad of
Scarbantia(PI. 27a) were also walled-up, and the statuesof the Triad at Savaria
also show tracesof destructionby force. It is, however,not always possibleto
decidewhetherthe smashedremainsof a sanctuarywere buried by Christiansin
order to get rid for goodof the idol in this way, or whetherthey were concealed
by pagansfor reasonsof reverence.But thereare collectionsof cult-objectsthat
have come down to us unharmed,as a treasurehidden away; in such casesthe
pagansobviously managedto hide away their cult-images and the sacrificial
implementsin time. Thesedestructionsand concealmentscannotbe dated; on
the whole the processwill havelastedroughly from the time of Constantineto
that of Julian.II 4
32 4
Under the Tetrarchy the governmentprobably took official steps for the
promotion of the old religion and cults. It has beenassumedthat it was for an
anti-Christian occasion that the priests of Dolichenus of Lower Pannonia
assembled. lIS But this assumption is untenablefor the reasonthat Dolichenus,as
we haveseen,disappearedfrom the late paganpantheontoo early. A sensational
new find proves,on the contrary,a deliberateand fully detailedpaganreligious
policy under the Tetrarchy.
At Brigetio a small group of graveswas found, dating fromthe beginningor
first half of the fourth century; most of them werein sarcophagi.In the centrally
situatedandrichestgravelay a manwho held a silver augur'sstaff (PI. 44a)in his
right hand. The very pretentiousgoldsmith'swork of the staff points to the end
of the third centuryas the earliestpossibledate: this dateof manufacturefor the
staff is likely to be significant. That it accompaniedthe augur to the grave-he
was evidently a municipal dignitary of Brigetio-is a striking illustration of the
religious circumstancesof the time. The lituus was intended to serve several
generationsof augurs;the man who took it to the gravewith him was clearly a
pagan,and his friends were evidently convincedthat he would be the last augur
of Brigetio.II6
Christianity and the persecutions

At the period when the augur of Brigetio could still practise his cult without
disturbance,therealreadyexistedin southernPannoniasmall groupsof people,
mostly Greek-speaking,who confessedthemselvesChristians.Their local tradi-
tion presumablydid not reachback much further than the middle of the third
century. The first Pannonianmartyr is indirectly establishedas the victim of
Valerian'spersecution,and underGallienussomeChristians,either during their
trial or during conversationswith fellow-believers,referredto their having been
broughtup in Sirmium as Christians.I I7 Someof the Pannonianmartyrs, there-
fore, were Christiansof the secondgenerationin Pannonia.As can be deduced
from their names,however,they were mostly immigrants-Orientalsor people
from the Greek-speakingprovincesof the empire.
The first certainindicationsof Christianity on the Danubecan be tracedwith
the aid of the Acts of the Martyrs. The earliestmartyr in our provincesseemsto
havebeenEusebius,bishopof Cibalae,whoseActs indeedhavenot comedown
to us, but whose memory has beenpreservedin local tradition. The Christian
lector Pollio of Cibalae,who was executedin 303, referredto his exampleamong
other reasonsbecausehis own martyrdomfell on the sameday as that of Euse-
bius (28 April). According to Pollio, Eusebiusmet his martyr's deathduring an
32 5
The final period oj prosperi~y
earlier persecution(persecutiosuperior), and this can most probably be identified
with the persecutionunder Valerian.1I8 It is not impossible that at that date
organizedChristiancommunitiesexistednot only at Cibalaebut in a numberof
other towns in south-eastPannonia-above all in Sirmium-andthat thesealso
had to suffer in Valerian'spersecution.
However,the life of the Christiancommunitieswas probablyundisturbedand
peacefulduring the last decadesof the third century,as can bededucedfrom the
writings of Victorinus, bishop of Poetovio,II9At this time therealreadyexisted
a few heresieswithin the communities,andit is not impossiblethat the Catholics
in Sirmium had a separatecommunityfrom that of the heretics.12o The appear-
anceof Christiancommunitieswith their own bishopsprobablyoccurredin the
period from Gallienus to Galerius; by 303, bishops and priests are already
attestedin severaltowns insouthernPannonia.The executionof the four or five
stonemasons, who later enjoyeda specialvenerationin Romeunderthe nameof
SanctiQuattuor Coronati, was probablyan isolatedinstance,not part of a system-
atic persecution.Sincethe Acts of thesemartyrsprove the personalintervention
of Diocletian, the martyrdom,which took place on 8 November,shouldprob-
ably be datedto the year 293 when Diocletian was presentin Pannoniaand-
specifically in November-madea journey along the limes. He ordered the
stonemasons,who worked in a porphyry quarry, to be executedbecausethey
refusedto producea statueof Aesculapius.121
The great persecutionof the year 303 throws a suddenclear light on the
Christiancommunitiesof our provinces.I22 And onething mustnot be forgotten
here: to all appearance the persecutionwas directedtowardsthe politically most
important area of Sirmium, where the procedurewas truly extreme.In other
parts of Pannoniaand Upper Moesia either no martyrs are known or we have
information only about the bishops.This meansthat in generalonly the most
prominentleadersof the communitieswere punished,and in someremoteareas
not eventhesewere consistentlypersecuted:we know almosttwo dozennames
of personslater reveredas martyrs from the surroundingsof Sirmium, whereas
no martyrs are known from the provincesof PannoniaPrima and Valeria, and
the statementson martyrs in Dardania,Dacia Ripensisand Dacia Mediterranea
contain hardly more than a few doubtful names and commemorativedays:
Florus andLaurusin Dardania,Hermes,Aggeusand Gaius on 30-3I December
in Bononia or Ratiaria; at Naissusthereare not even martyrs'names.The only
martyr in the provinceof Saviawas the famous Quirinus, bishop of the provin-
cial capital at Siscia. Quirinus was executedin Savaria,presumablyin order to
intimidate the Christiansof that town.
The possibility, however, cannot be excludedthat the transmissionof the
326
namesof martyrsandthe survival of their cults dependon the later history of the
Christian communitiesin question. Sirmium preservedits role for a very long
time after the collapseof the westernempire; in the gradual emigration from
Sirmium, the relics of martyrs and their cults were taken along and a new cult
establishedin the new home. One of the most important saints of the Greek-
Orthodox and Slav churcheswas the Sirmian Demetrius, whose relics were
transferredto Thessalonica.The emigrationto Italy in the beginningof the fifth
centuryplayeda decisivepart in the survival of a numberof other martyr-cults,
for instancethat of Quirinus and the four Coronati.But the martyrsand cults of
some towns in which their Christian communities died out early in the fifth
centurycouldeasilyhavebeenforgotten.Nevertheless,the northItalian churches
havepreservedthe memoryof a long list of Danubianmartyrsandcommemora-
tive days, and since thesewere for the most part Sirmians,it is quite probable
that most of the martyrs came from Christian communities in and around
Sirmium. But there remains the possibility that the martyrs of less important
centresin our provinceswere much more easily forgotten.
The long list of Christians executedin Sirmium contains some namesthat
receivedespecialglory in the cult of the medievalchurch, aboveall Demetrius,
deaconof Sirmium (the Saint Demetrios(Dmitrij) of the EasternChurch)and
Irenaeus,bishopof Sirmium; the gardenerSynerotasand FortunatusandDona-
tus had a specialcult in Sirmium; their placesof rest becamethe object of pious
veneration and focus of a large Christian cemetery there.123 Synerotaswas
presumablynot executedin the sameyear as the others;his commemorativeday
is 23 Februarywhilst most of the others died their martyr's deathin March or
April. The commemorativedays, which are concentratedin the spring, allow a
tentativereconstructionof the progressof the persecution.In 303 all was quiet at
Sirmium at first. At the endof March andthe beginningof April the headsof the
communitiesat Sirmium and Singidunum were arrestedand executed:these
were Bishop Irenaeusand the deaconsor priests Fortunatus,Donatus,Deme-
trius, Romulus, Montanus (togetherwith his wife Maximilla) and presumably
many more, for instancethe Seven Virgins whose individual names are not
known. At the end of April the persecutionsextendedto the neighbouring
towns: at that time Pollio, lector of the Christiansat Cibalae,was executed,and
probablynot much later the bishop of Siscia, Quirinus, was arrested.Tradition
speaksof Montanusfleeing from Singidunumto Sirmium and being therefore
convicted there. But it is probable that Christians from neighbouring Singi-
dunum were brought to Sirmium for trial and judgment.
Possiblyfor different reasons,Bishop Quirinus was broughtto Savaria.After
having beenarrestedon his flight he was led 'throughthe towns', and the trial
327
was held at Savaria,andtherein the theatre,that is, with full publicity. From the
Acta we learn that at that time Savariatoo alreadyhad a Christian community;
but of Savarianmartyrs tradition knows nothing. The ostentatiousmanagement
and the delay in the criminal proceedingscan therefore be explained in the
following manner:the Christiansof the provincesof Saviaand PannoniaPrima
were not so strictly and consistentlypursuedas thoseof PannoniaSecunda,and
the exampleof Quirinus was intendedto suggestthe possibility of future perse-
cution. On 8 June Quirinus was flung into the small river flowing through the
town of Savaria. In the place where his body came to land an oratory (locus
orationis) was later built, and under Constantinehis relics were depositedin the
basilica erectedfor him.
In Sirmium the persecutioncontinued longer, presumablyuntil 304. The
commemorativedays of the remaining martyrs, among them the gardener
Synerotas,are spreadover the monthsJuly to February;nothing is said, how-
ever, about their being dignitaries of the church hierarchy, and many women
were also includedin their number.This was the stagein the persecutionwhen
laymen too were falling victims. It is very probable that smaller, more local
persecutionscontinuedto occurunderLicinius too; for instance,it was probably
under him that Hermyle and Stratoniceof Singidunumwere thrown into the
Danubethere. The martyrdomof Synerotas,too, can be datedto the later years,
c. 305-7.
As we have seen, the persecutionsunder the Tetrarchy suggestthat in the
larger centresof our provinces organizedChristian communitiesexistedunder
bishops.The sameassumptionappliesalso to towns which had neither martyrs
(or more correctly martyrs that were later venerated)nor Christiansmentioned
for any otherreasonin our sources.The next occasionwhenwe learnsomething
about the organizationof the churchon the Danubeis the council of Nicaea,at
which the Christian communitiesof Pannoniaand Upper Moesia were repre-
sented.The Danubianbishopsnamedamongthe fathers of the council mostly
signed as representativesof a province-DomnusPannoniae, Dacus Dardaniae-
and, consideringthe distancefrom the Danube to Asia Minor, this possibly
meansthat they attendedas the emissariesof a whole province, since not all
bishopsof the province could take part in the council.
We learn somedetails of the diocesesin our provincesonly from 343, when
some otherwiseunknown bishopstook part in the council of Serdica.At this
council, too, not all PannonianandMoesianbishopswere present;the majority
were bishops of towns situated nearestto Serdica, such as Horreum Margi,
which is attestedas a see onlythis once. There are, in fact, hardly any towns in
Pannoniaor Moesia whose fourth-century lists of bishops are handeddown
328
without gaps.But it is probablylegitimateto assumethat by the first half of the
fourth centuryall towns (civitates) in the Danubianprovinceshad bishops,even
where these arenot namedin the sources.The following towns of our pro-
vinces are attestedas bishops'sees:
PANNONIA PRIMA: none.
VALERIA: none.
SAVIA: Siscia, Jovia.
PANNONIA SECUNDA: Sirmium, Mursa, Cibalae.
MOESIA PRIMA: Viminacium, Singidunum,HorreumMargi, Margum.
DACIA RIPENSIS: (westernpart) Ratiaria, Aquae.
DACIA MEDIl'ERRANEA: (westernpart) Naissus,Remesiana.
DARDANIA: Scupi, Ulpianum.
A complete bishops' list we have only for the fourth century and only for
Sirmium. Of the remainingseesonly somebishopsare known by name;Aquae
and HorreumMargi are mentionedmerely in the list of the council of Serdica,
and the bishopric of Margum is attestedonly by the fact that a bishop, not
mentionedby name, had an embarrassingexperiencewith the Huns.I24 The
bishop of Jovia appearsonly once, on an inscription,I2S and lastly, certain
Moesianseesare attestedonly for the fifth or the sixth century.By andlarge our
knowledgedependson a variety of chances:the geographicalsituation of the
town in question,the part individual bishopsplayed from time to time in the
general history of the church (for instanceValens of Mursa and Ursacius of
Singidunum,etc.). Taking all thesechancesinto consideration,it would not be
right to maintainthat there were no bishopsin PannoniaPrima and in Valeria;
we must rememberthat in other provinces,for instanceMoesia Prima, almost
everytown is attestedas an episcopalsee.The evidenceof early Christianmonu-
mentsso far discoveredsuggestsa bishopricat any ratein the provincial capitals
of Savariaand Sopianae;and with the organizationof the early churchin mind
we can say that it is quite probablethat therewere bishopsin the remaininglate
Roman civitates as well.
We are, of course, best informed on the church of Sirmium, although its
bishopsdid not playapart of first importancein the dogmaticdissensionsin the
middle of the fourth century. The leading roles were played by the bishopsof
two towns near Sirmium: Valens, bishop of Mursa, and Ursacius, bishop of
Singidunum, were the leaders of the Arian party in the western part of the
empire, and exerteda decisive influence on the governmentof ConstantiusII
which was well-disposedtowardsthe Arians. With their appearancethe ecclesi-
astical history proper of our provinces begins; until the disintegrationof the
329
Romanadministrationand provincial organizationthis was determinedby the
controversybetweenthe Catholics and the Arians. The bishops of Sirmium
were as dependenton the powerful influence of Valens and Ursaciusas most of
the bishopsof our provinceswere; the only exceptionwas Photeinos,bishopof
Sirmium, who was repeatedlyin exile or being reinstated: he representeda
hereticaltrend of his own.
The activity of Valens and Ursaciusbeganin the 330S when they madetheir
appearanceas eagerfollowers of Arius (who had beenexiled to Illyricum) and
succeeded in obtainingthe bishopricsof Mursaand Singidunumrespectively.At
the same time they succeededin sending the Catholic bishop of Sirmium,
Domnus, into exile, and in establishingin this important bishopric the Arian
Eutherius.Valens'attempt,in 341, to becomebishopof Aquileia failed, andthis
incident causeda temporary decline in PannonianArianism. The bishops of
south-westernPannoniaseem to have been under the influence of the north
Italian churches,always on the side of Catholicism, and in the council of
Serdicathe Moesian bishops also took sides with the orthodox Osius. Even
Eutheriuscameover to the Catholic side, and this left Valens and Ursaciusno
alternativebut to professCatholicismthemselves.
In the sameyear, 343, however, Eutheriusdied, and the new bishop of Sir-
mium, Photeinos, only weakenedthe unity which had been so difficult to
establish;as early as 346 Valens and U rsacius,who most of the time travelledin
affairs of church politics, appeared againstPhoteinosin Milan. A synod in
Sirmium not much later condemnedPhoteinos,and in 351, in the presenceof
the Emperor ConstantiusII another synod was convokedat Sirmium which
sentPhoteinosinto exile. The new bishop,Germinius,was an Arian, for Valens
had in the meantimesucceededin establishinghis influenceover the emperorby
prophesyingto Constantiusat his prayershis coming victory over Magnentius
nearMursa. As a church-politicianBishop Germiniuswas hardly more than the
mouthpieceof Valens.
The last years of ConstantiusII were under the influence of Arianism, but
alreadyby 358 therewere indicationsthat westernCatholicismwas on the point
of regaining Sirmium. Two synods at Sirmium (357, 358) showed Arianism
alreadyin retreat,andat Rimini ValensandUrsaciusweredefeated.The decisive
impulse camefrom north Italy, particularly when Photeinoswas recalledunder
Julian and the Sirmian church thus fell victim to a different type of heresy.
Although in 365 the Arian Germiniuswas allowed to recoverhis see,from then
onwardsValens and his irresolutefollower Germinius gradually lost ground;
and whenValensand Ursacius,by now probablyold men, disappearedfrom the
scene,Ambrosian Catholicism succeededat last in establishinga bishop of
330
VlVAS~ R<iIOD S~CI

GRAEC~EXREG~ DI~
zIX~ N~L~ET R~ER
T NI~A~I~E
T NI~A~I~ENI~A~I~E
(tL
TT NI~A~I~E
NE §nVIII D NIC
C
§nVIII
EM~ N~AR~E~FIL~
I ECI T~
§nVIII
Figure 55 Christian tombstoneof Aurelius Iodorus, a Greek from Laodicea,found
at Savaria(HungarianNational Museum)
orthodoxviews in Sirmium. Under Valentinianthe church of Sirmium with its

bishop Anemius was already under the influence of north Italian Catholicism.
An eventwhich would have beenimpossibleonly a quarterof a centuryearlier
happenedin 378: western Roman bishops assembledin Sirmium and unani-
mously condemnedthe eastRomanArian heresy.
It is to be hopedthat this survey,despiteits brevity, has shownthat any part
playedin generalchurchhistory by the Christiancommunitiesof Pannoniaand
Upper Moesia was due to their proximity to the imperial court at Sirmium and
the opportunity this afforded for involvementin high politics. The bishopsof
towns that lay further from Sirmium either played no part at all or entered
church politics only as secondaryfollowers of this party or the other; and this
may havebeenthe reasonfor their being namedonly in exceptionalcasesand by
chancein the sources.
After all that hasbeensaid,it would seemidle to ask whetherthe Christiansof
our provincesprofessedthemselvesArians or orthodox. Not even the leading
33 1
The final period oj prosperi~y
personalitiesof the communitiescreatedor representeda dogmatictrendof their
own; on the contrary,a sometimesunattractivemanoeuvringwas characteristic
of the most importantbishops.The Arianism of a Valens or a Ursaciusseemsto
have been mainly dependenton the general situation of ecclesiasticalpolitics
and not on the opinion of their own community or its priests. Although they
were the most eagerexponentsof easternArianism in the west, their influence
and their successesrestedall too much on the favour of the Court. The broad
mass of believers were presumably little affected by politically conditioned
dogmaticsubtleties,and were all the more readyto leave controversiesto their
priestsand bishops,sincethey were mostly still involved in the simple dilemma
of Christianity versus paganism.That Arianism was able under such circum-
stancesto gain the upper hand was probably mainly due to the fact that the
earliest,and thereforedogmaticallybest orientated,Christianswere, as we have
seen, Greeks and Orientals, whilst it was only in the first half of the fourth
centurythat the native populationbeganto be converted.The oriental compon-
ent in the Christian communitiescan be demonstratedeverywherewhere early
Christian inscriptions have come to light (Figs 56, 57), and even the language
of many early Christian grave-inscriptionsis Greek,I26whereasthe Greek ele-
mentin the epigraphicremainsof earlierperiodswas very small. Whether,in the
Arianism of the Pannonianbishops,we are observingtactics of an Illyrican kind
for securingDanubianinfluencein the policies of the westernchurch,cannotbe
proved though it is not improbable.
Early Christianmonuments,of course,can only prove the existenceof Chris-
tians: they cannotindicatethe proportionof Christiansto pagans.But thereare
some monumentswhose interpretationis questionable.A number of ground
plans, the results of old excavations,have been interpretedas Christian cult-
rooms,andin somecasesmodernre-excavationhasprovedthat theseinterpreta-
tions are incorrect.Apsidal rooms are often consideredto be Christianbasilicas,
althoughthe groundplan itself admits other interpretations.For instance,it has
recentlybeenshownthat the so-calledbasilicaof Quirinus at Savariawas in fact
the Aula of the imperial palaceandnot the resting-placebuilt underConstantine
for this bishop of Siscia.I27Despitethis, the Christian monumentsof Pannonia
are very numerousand varied, particularly in the better-explorednorthernpart
of the province, and a similar wealth of monumentscan well be expectedin
UpperMoesiatoo, whereexcavationsat Ulpianum (PI. 38b) and Naissussuggest
large Christian communitieswith extensivecemeteries.
Indeed, the most notable remains of the Christian communities are their
cemeteries.It is extremelylikely that in the fourth century Christiansinstituted
cemeteriesof their own, separatedfrom those of the pagans,at least in those
332
A
VlVAS~ R<iIOD S~CI
GRAEC~EXREG~ DI~
zIX~ N~L~ET R~ER
T NI~A~I~E (tL
NE §nVIII D NIC
C
§nVIII
EM~ N~AR~E~FIL~
I ECI T~
§nVIII
A
Figure}6 Christian tombstoneof Artemidora buried near St Synerotasat Sirmium
settlementswhere they were numerically powerful. On an inscription from the

Christian cemetery of Savaria a custor rymiteri is named,uS and in all larger
cemeteriesso far investigatedit could be establishedthat the centre was the
cult-chapelof a deceasedChristian who was particularly reveredin the com-
munity. A few written sourcesmention martyrs' resting-places,and epigraphic
evidence129 makesit certain(Fig. 56) that the focus of the Christiancemeteryat
Sirmium was the burial-placeof the martyrs Synerotasand Fortunatus,who are
of courseotherwisealso well attested.
The centre of the Synerotascemeteryat Sirmium was a Cella Trichora; this
type of cemetery-chapelhas also beenfound in Sopianaeand Aquincum (Fig.
53).130 The remaining chapels in undoubted Christian cemeteriesare distin-
333
Figure J7 Paintedvault of the cella memoriaeat Pees(Sopianae)
guishedby their variety.I3I A simplified variant of the Cella Trichora was the
chapelwith two apsesin which, for the most part,sarcophagiwere placed;this
type is attestedat Ulpianum and in the cemeteryof Tricciana (Sagvar).I32The
sarcophagiwere either buried in the ground or they stood on the floor of the
chapel (PI. 38b); for the most part they were simply executedin the so-called
housetype, without decoration.The Belgrade sarcophagus, re-workedfrom an
earliersarcophagus with the representationin relief of the GoodShepherdandof
Jonah,is unique,133as is also the sarcophagus from Szekszardin the provinceof
Valeria,134though this carries Christian symbols only on its lid. In the grave-
chapelsthe deadweremostly buriedbelow ground,quite often in a subterranean
grave-chamber.These grave-chamberswere often painted, in a rather unpre-
tentious,simple, but very engagingand straightforwardstyle. The best-known
monumentsof this kind are Grave-chapelI at Sopianae(PI. 4ob) and the closely
relatedexampleat Naissus,both of which showthe Chi-Rho monogrambetween
PeterandPaul in the main scene;135 this iconographicdesignis alsofound in the
apse-mosaicsof basilicasin Rome, datablenot much later. At Sopianae,there
are also portraits(?) in medallions and scenesfrom the Old Testament(e.g.
Jonah,Noah, etc.) representedon the vault (Fig. 57). A further commontrait
334
on murals at Naissusand Sopianaeis the fence of a garden(PI. 4oa), intended
to symbolize the garden of Paradise-orso it is interpreted, presumably
correctly. In a simplified mannerthe motif of the gardenis painted in a very
small stone-builttomb, also in Sopianae.I36
Exceptionally,grave-chapelsalso had mosaicfloors, for instanceat Ulpianum,
but otherwisethey were probably decoratedin a rather simple fashion. The
larger chapelscontainedmany burials, the dead being walled in next to each
otherin the chamber.In the largestchamberthat hasso far cometo light, found
in Sopianae,there was room for twenty dead.I37 Apparently there existed no
connectionbetweenthe size of the chapelsand their artistic or architectural
equipment. The painted chamberI at Sopianaehad a rather modest super-
structure, whilst the unique Cella SeptichoraI38 at Sopianae(Fig. 53) had no
subterraneantomb. Modest chapelswith one apse or even with only a rect-
angulargroundplan areattestedin manycemeteries,but in the hugelate Roman
cemeteryat Naissuscemetery-basilicas of oftenconsiderablesizewereuncovered.
It is quite probablethat the crypt of the eleventh-centuryCathedralof Pecs
(Sopianae)was an early Christian cemetery-basilicatoo, especiallyas it has a
basilicanplan with naveand two aislesand as the cathedralis built in the areaof
the early Christian cemetery-whichitself is presumablynot accidental. The
Cella Trichora of Sopianaewas rebuilt in about the ninth century and its walls
repainted.
The early Christiangrave-inscriptionsofPannoniaandMoesiaarealwayssmall
tablets,which were built into the walls of grave-chapelsor servedto close the
loculi. Thesealone are sufficient to prove the rich architecturalachievementof
the early Christiancemeteries:in someof them suchgrave-tabletscameto light
in greatnumbers,for instanceat Savariaand Sirmium. And this is evidencethat
cemeterieswith rich furnishings plentifully suppliedwith buildings aboveand
below ground are Christian. There are of coursea large number of extensive
fourth-centurycemeterieswhich althoughfairly rich in grave-goodsare fairly
uniform in burial-rite and decidedly poor in architecturalelaboration;pagan
burials are probably to be soughtin thesepoorer surroundings.
The centresof Christianity, however,canprobablybe lookedfor in the larger
settlementsand in the large estates.This is suggestednot only by finds of
Christiangrave-goodsbut alsoby the cult-buildingshithertocertainlyidentified.
As we haveseen,the cemeteriesthat can certainlybe establishedas Christianare
municipal cemeteries,or at least those belonging to large settlementsand to
fortified centresof imperial latifundia. It is in such places that Christian cult-
buildings are to be expected,and in addition in some larger villas. Christian
basilicas (Fig. 53) have beenproved at Ulpianum (PI. 38b), Remesiana-both
335
episcopalseesin the fourth century-Aquincum,and at Fenekpuszta(PI. 3Sa),
andalsoin the areaof a late Romanvilla north of Lake Pelso,wherebrokentiles
with the Chi-Rho monogramsuggesta Christiancult.I 39 Marble altar-slabshave
cometo light in the villa of Donnerskirchen,at the villa of Orvenyes-whichhas
also yielded other piecesof architecturalsculpturefrom a cult-room-andfrom
Csopakin the vicinity of Orvenyes;140 the last two find-spotsare situatedon the
north shore of Lake Balaton (Pelso), where great estateshad long been the
predominantagricultural type.
Art in the late Romanperiod

If Christianityappearsto some extentas a religion linked to urban centres,the
administrationand ownersof estates,this is evenmoreso with late Romanart in
the Danubianarea. The first two periods of prosperityunder Romanrule had,
as everywhereelse,produceda provincial Romanculture of a specialcharacter,
traceablefor us in a local style of sculpturein the gravemonuments,in the minor
arts partly derived from pre-Romantradition, in a few local cults and in other
similar characteristics.Although this provincial Roman civilization was the
culture only of the ruling upper class of provincial society, this class had a
sufficiently numerousmembershipand was presentpractically everywhere,as
eithera military or a municipal elite, or in earlier times, as a tribal aristocracy.In
times of peacethis class had richly ornamentedgrave monumentsset up with
inscriptions in Latin, donatedcult-buildings, cult-statues,reliefs, votive gifts
and altars, adaptedits way of life more and more to Romanstandards,and was
keenin all its lasting manifestationsto appearas a Romanupperclass.Thoughit
declinedin numbersand in wealth under the Tetrarchythe class continuedto
exist and left behind many traces of its earlier outlook. But soon it reached
completestagnation,and this has to be attributedto far-reachingsocial changes
which we havealreadydescribedas far as possible.Theartistic monumentsof late
Romantimes stemnow from a new aristocracy,which was in part an official class
tied to the administrationof the empireand in part a necessarilysmall group of
great landowners.It was, on the one hand, numerically much weakerthan the
earlier military-municipal aristocracy, and therefore was not able to sustain
a local artistic production; but on the other, as part of the imperial aristo-
cracy, it was rich enoughto patronizeimperial art and to cultivate centresof art
andartistsbeyondthe province.Early Christiangrave-inscriptionswhich for the
most part hadbeensetup by the middle classeswerefor that reasontabletsof no
artistic pretension,lacking plastic executionand carvedwith clumsily chiselled
letters: the customershad to apply to local stonemasonswho were no longer
33 6
accustomedto plastic works that a stonemasonof Severantimes would still
have beenable to executewithout difficulty. The new aristocracy,on the other
hand,sentfor mastersfrom Aquileia whenthey wishedto adorntheir villas with
mosaics;they acquiredworks in silver (PI. 4 5a) or in gilded bronzefrom the
important workshopsand had sculptorsinvited to attendthem.
But we must not underratethe proportionof late Romanartistic monuments
in Pannoniaand in Upper Moesia which are attributablenot to this aristocracy
but to the emperorhimself, who was either the owner or the donor of these
works of art, and often the object of the representationtoo, as was the casewith
the imperial statuesof Naissus,for example.The evidently very rich sculptural
decorationof the villa at Mediana, or the statuesin porphyry found here and
there-andsignificantly in Gamzigradin particular-werecommissionedfor a
high official if not the highest; finds like the gilt-bronze portrait of a young
emperorfrom Sopianae,14 1or the statecameofrom Kusadak142 in north-western
Moesia evidently derive from court workshops-andfrom the view-point of
local art and culture it is irrelevant that some of these courtworkshopswere
centred at Sirmium and Naissus. The gilt parade-helmetsof high-ranking
officers found in the fort oppositeAquincum and at Berkasovo,143not far from
Sirmium (Frontispiece),are the productsof factories listed in the Notitia Digni-
tatum. The silver dishesfrom various late hoardswere producedin workshops
at Naissusand were probably intendedto contain imperial gifts and presenta-
tions for dispatchto personsof high rank.144The portrait-bustfrom Sopianae,
already mentioned,can also be interpretedas an imperial present,being fixed
either to the tip of a consular sceptreor to the middle of a dish. The richly
ornamentedsilver tripod from PoigardP45in Valeria (PI. 45a) could equally
well have beenfound in one of the famous late silver-hoardssuch as those of
Augst or Traprain.
Even someof the less valuableproductsof the appliedarts shouldperhapsbe
attributedto central workshops.Possibly the most characteristicgroup of late
Roman small artistic objects in Pannoniais the bronzemounts of small boxes
with ornamentationin chasedrelief (PIs 44b, C),146 The themesare very varied,
but are related to imperial art and have a certain official character.Thesesmall
boxes,too, canbe interpretedas official presents-of sofar unknowndestination.
Lastly, possibly the most striking examplemust be mentioned:the technically
skilful, complicatedcross-bow brooches so uniform in the manner of their
production,and in their ornamentation,that they must certainly originatefrom
state-ownedfactories. Among them are gilt, nielloed, even gold examples,
notableproductsof a very exactinggoldsmith'sart. From an earlierperiod, too,
there exist brooches that were the products of a highly developed artistic
337
industry.But they arelessuniform than the late Romancross-bowbroochesand
for typological reasonsmust be attributedto local masters;for instancethe so-
called Norico-Pannonianwing-broochesof the native Celtic population are
notableproductsoflocal metal-workingandoflocal taste.But thereareno types
of finds and groupsof finds of late Romandatewhich show local characteror a
locally limited distribution. Costumeand way of life had becomelargely stan-
dardized,not only in high social circles, which though weak in numberswere
closely interlinked with the aristocracyof the empire, but also amongthe com-
mon people, whose way of life becameuniform presumablyas the result of
poverty. Social changein the late Romanperiod resultedin the former middle
classes,composedof the military and the municipal aristocracy,being either
promoted into the imperial aristocracy or-presumablythe more common
alternative-sinkingback into the commonpeople. This polarizationfinds its
expressionin the history of the settlements,in the growth of large estates,in the
materialof the small finds, in burials, in the arts and in the last resortin religion
too. The processmust be summarizedand emphasizedonce againat the end of
this chapter,sinceit hadfar-reachingconsequences for the fate of our provinces
after the end of the fourth century. Roman culture and presumablythe Latin
languageas well were confinedto thosewho by office or by wealth wereclosely
connectedwith the governmentand with the administration;they no longer
formed a military-municipal social class mediating between superiors and
inferiors, but a classthat was ratherisolatedfrom the people,andwhich was not
strong enoughto maintainitself underthe changedcircumstancesafter the end
of the fourth century,in the way in which the senatorialestate-ownersof Gaul,
Spainand Africa were able to maintainthemselves.The specialsituationof our
provincesas frontier-provincesmay have beenresponsiblefor this: the incor-
porationof massesof barbariansfrom the left bank of the Danubeaffectedthe
Danubianprovincesmost of all, and the largestestateshere probably belonged
to the emperor;particularly so sincetheseestateshad comeinto existenceonly
lately, as the result of politico-military changesand of a systematiceconomic
policy.
338
Chapter 10
The beginning of the Dark Age
In the autumnof 376 hordesof Goths underthe leadershipof Fritigern crossed

the lower Danubeand were given the receptiointo the empireby aJoedus,which
allowed them to settle in the provinces of Dacia Ripensis, Moesia Inferior
(Moesia Secunda)and Thrace.This dateis a turning-pointin the history of the
Danubianand Balkan provinces,since after it barbariangroups never ceased
to be found within the frontiers of the empire. Late Roman historians fully
appreciatedthe significance of this turning-point; Jerome, by birth a Dal-
matian, starts the final misfortunesof the Romansat Illyricum at this date,I
contemporaryand later writers do not tire of speakingof the fall of Illyricum,
and a chronicler of the easternempire took the year 377 as the beginning of
the rule of the Huns in Pannonia,2 and not wholly without reason,as will be
seen.
The Gothswho hadbeensettledon imperial territory startedthe sameyear to
becomerestless,and plunderedfar and wide in the provinces that had been
assignedto them. The immediatecauseof theseraids was that the Goths were
not, or not regularly, given the provisionswhich they hadstipulatedand which
they shouldhave obtainedasJoederati. The situation was further aggravatedby
the appearance of a new group of Gothsunderthe leadershipof Farnobiusfrom
across the Danube. The first Roman counter-measure,taken in 376, merely
succeededin diverting the plunderingGothsfrom Gratian'spart of the empire.
This was due to Frigeridus(possiblya former dux of Valeria), who, on Gratian's
orders, appearedin Illyricum with western,including Pannonian,forces. The
next year, Gratian sent more troops under Richomeresto Illyricum, and these
made contact with the easterntroops of Valens. In the meantime,however,
339
The beginningoj the Dark Age
there had also been an invasion of Taifali from Trajan's Dacia into Dacia
Ripensis;they placedthemselvesat the disposalof Farnobius.Frigeridus suc-
ceededin defeatingthem south of the Haemus(in the Balkan mountains),but
this did not changethe fact that the Gothswerein rebellion on imperial territory
and neither westernnor easterntroops could either throw them back over the
frontier or force them to be peaceful.
In spring 378, therefore,both emperorsset out for the Balkan peninsulawith
their field-armies. Valens came from Antioch and Gratian from Trier. For
Gratian the organizationof the campaignwas so urgent that he went by boat
down the Danubeand stoppedonly at Sirmium for a short halt. From Sirmium
he also probably went by boat, and did not set foot on land before the eastern
edgeof DaciaRipensis,wherehe found himselffacedby a group of Alans. Near
CastraMartis he suffereda minor defeatat their hands,which howevercontri-
butedconsiderablyto his decisionnot to advanceany further. Valensis attested
in Constantinopleat the end of May,3 and it will have beensomewhatlater that
he appearedin Thrace.
The group of Alans that opposedGratian near CastraMartis belongedto a
Gothic-Alan group whose leaders were the Goth Alatheus and the Alan
Saphrac.A band of Huns which had desertedthe main body of Huns also
enteredthis alliance; a circumstancewhich was to be of a certainimportancein
the event. This mixture of peopleunder Alatheusand Saphracenteredthe em-
pire probably in 377 and at first remainedin Dacia Ripensis. In 378 the two
chieftains got in touch with Fritigern and contributed considerablyto the
catastrophenear Adrianople on 9 August. After the battle, like the other bar-
barian groups, they went on raids in the Danubianand the Balkan provinces,
and finally cameto Pannonia.4
Gratian remainedat Sirmium during the winter of 378-9 where he created
Theodosiusan Augustus on 19 January.sSoon afterwards he left Illyricum,
leaving the problem of the Goths to the new emperor. Theodosiuswent to
Thessalonicawhere he made preparationsto fight the barbarians.Only after a
numberof measureshad beentakendid he leaveThessalonicafor the purposeof
clearing the valley of the Axius and Dardania; how far north he managedto
advance,we do not know.6 The chronologyof this year is very confused.We
are informed of a military enterprisethe leaderof which was the dux etpraesesof
Valeria, Maiorianus,who was soon afterwardsappointedmagister.7 Possiblyhe
fought againstthe groups led by Alatheus and Saphracwhich had penetrated
into Pannonia.A victory over Goths, Alans and Huns was announcedin Con-
stantinopleon 17 November379,8 while Theodosiushad suffereda numberof
smaller defeatssouth of Pannoniaat the hand of Fritigern.9 The Alans, Goths
340
The beginningof the Dark Age
andHuns, however,who were defeatedin the autumnmay be identifiedwith the
peoplesof Alatheusand Saphrac.The victory over them (? in Pannonia)could
have no wider results becauseof Theodosius'lack of success,and we hear, in
fact, that the towns of Pannoniahad to live throughdifficult times in the winter
of 379-80, and some settlementswere even destroyed.The commanderof the
Pannoniantroops, Vitalianus, was powerlessagainstthe barbarians,who were
attackingon a wide front. 10
Gratian had therefore to appear againin Illyricum. He sent a number of
troops aheadof him, who undertookclearing-upoperationsin Dacia Mediter-
ranea and Thrace,II and at the end of the summer came to Sirmium, where
Theodosius,who had beenimpededby illness, also arrived not much later. At
the meeting of emperorsin Sirmium, which was probably regularly repeated
during the following yearsat either Sirmium or Viminacium, someprinciplesfor
regulatingthe situationin Illyricum, which had fallen completelyinto disorder,
could be worked out: in the following years negotiationswere set going with
the Goths of Illyricum, leading, in 382, to a new treaty with the Goths, accord-
ing to which they were given extensiveareasfor settlementin the provincesof
Dacia and Thrace. In 380 the peoplesof Alatheus and Saphracwere settledin
Pannonia,presumablyin the areasof which sincetheir plunderingsin the winter
of 379-80 they were the de facto masters.12
We haveno informationon the sizeandpositionof the areasdesignatedby the
treaty with Alatheus and Saphrac.The only fairly secureclue is given by the
grave-inscriptionof a bishop, who was buried in Aquileia approximatelyat the
beginning of the fifth century. In another source, this Bishop Amantius is
attestedin 381 as bishop of the small Pannoniantown of Jovia in the valley of
the Drave, eastof Poetovio,and his grave-inscriptionpraiseshim among other
things for having 'emigrated'(presumablyfrom his native town of Aquileia) to
Jovia, where he undertook a blessedactivity among the 'two peoples' and
developedgoodconnections'with two duces'.Whetherthe two peopleswere the
Romansof Jovia andthe barbarianfoederatior whetherthey werethe subjects of
the two chieftains,cannotbe decided;it is certain,however,that thegeminiduces
must be Alatheus and Saphrac.I3 From this we may infer that thesebarbarian
chieftains lived in the diocese of Amantius, or at least in the vicinity of his
diocese.Jovia lay in the valley of the Drave,of which the largestPannoniantown,
Mursa, was troubled by barbariansin the winter of 379-80. Jovia was, more-
over, situatedin the province of Savia, on the southernboundaryof which lay
Stridon, the birthplace of Jerome,which also underwenta barbarianattack in
379-80. A barbarian tribe-called antiqui barbari-who were possibly the
descendantsof thesefoederati, are mentionedin the province of Savia in the
341
sixth century.I4 Thus it is suggestedthat the settlementarea of the joederati
of Alatheus and Saphraclay in the valley of the Drave in the province of
Savia.
In the meantimethe situation beyondthe Danubehad changed,too. We do
not know exactly, of course,what happenedduring theseyearsin the approaches
of PannoniaandMoesia,but Ammianusmentionsunderthe year 378 that all the
peopleson the left bank of the Danube,from the Marcomanniand Quadi to
those living by the Black Sea, had started to move, and, driven from their
homes,were threateningthe limes.Is Presumablythey wanted to be given the
receptio-butthe receptionof Fritigern in 376 did not serveas a good precedent
to inducethe Romansto acceptfurther groupsof barbarians.Hoards,hiddenin
the ground in the Hungarian Lowlands around 375,16 suggestunrest in the
country of the Sarmatians;the Taifali, under pressurefrom the retreat of
Athanarichinto the southernCarpathianmountains,broke into Dacia Ripensis,
as we have seen,and in 381 evenAthanarichhad to ask for the receptio,presum-
ably as the result of Fritigern'sGothshaving had to retire over the Danubeinto
the old Dacia a short time previously,I7The reasonwhy we cannotestablishthe
effects of unrest beyond the frontier on the limes of Pannoniaand Moesia is
probably that the attention of contemporarieswas too much focused on the
situationinside the empire. An invasionof Sarmatiansoccurredat an unknown
datebefore384,18but nothingis otherwiserecordedof fighting on the limes; this
is partly due to the fact that pressureby the Huns at that time only affectedthe
lower Danubeand becausethe situationin the approachesto the limes had been
relieved by the emigration of large massesof Goths. Gothic pressureon the
Sarmatiansthereforedecreased,and the disturbancesof c. 375-80seemsoonto
have beensettled.
Thereare, indeed,someindicationsfor a certainamountof consolidationon
the middle Danube in the 380s. Bishop Amantius at Jovia has already been
mentioned.St Ambrosereports that in north Italy Pannoniancorn was traded
for wine in 383,19and Valerius Dalmatius,a governorof the Gallic provinceof
LugdunensisIII, at the earliest towards the close of the 380s, retired to his
estatesin the southof the provinceof Valeria.20 Since,however,the provinceof
LugdunensisIII was createdby the usurperMagnusMaximus, and its governor
was thereforeprobably one of his supporters,we must supposeeither that the
area controlled by Maximus included Valeria for a time, or-less probably-
that the former supporterof Maximus was exiled to Valeria. It is not impossible
that MagnusMaximus had considerableinfluencewith thejoederatiin Pannonia,
since in the same year as Maximus appearedin Pannonia,and in Savia in
particular,barbariansplunderedthe province.21 And Maximus got in touchwith
342
the Goths on the lower Danube, which enabledhim to hinder Theodosius'
advanceagainsthim.22
The towns of Savia, among them the important baseof Siscia, fell into the
handsof Maximus in 387. Theodosiushoweverwas able to takethe field against
him only in the summerof 388. He advancedfrom ThessalonicathroughDar-
daniaandMoesiaI to Pannonia,and defeatedthe usurper,first near Siscia, then
nearPoetovio.23 A factor of importancefor an estimateof future eventsmust be
mentionedhere:the mint of Siscia, which with its four ojjicinae was the largestin
Illyricum, ceasedproductionin this time.24 The greatmints of the west (Arelate,
Lugdunum,Treviri) ceasedproductionmostly towardsthe end of Theodosius'
reign, andat that sametime the other Pannonianmint at Sirmium (which was of
little importance) closed down.2s The closing of the western mints c. 395 is
evidently the effect of general economic and political conditions-in the last
resort the changeto an economybasedon natural produce; the drying-up of
monetarycirculation under and after Theodosiuscannot, therefore,be simply
connectedwith decayof 'the Romanway of life'. The closing-downof the mint
of Siscia broadly, of course, belongs here too, but the date, almost a whole
decadeearlier, demandsspecial reasons.In my opinion two motives must be
consideredhere:first theJoederatiof Alatheusand Saphrac,as we haveseen,were
settledprecisely in the province, or preponderantlyin the province, of which
Siscia was the capital. In this way the surroundingsof the Moneta Sisciensishad
becomea dangerarea,andit presumablyseemedadvisableto the governmentto
shut down this important mint. Secondly,it has recently becomeincreasingly
clear that it was not under Theodosiusthat money circulation in Pannonia-
aboveall north of the Drave-ceased, but earlier, c. 375: in other words to all
appearanceconsiderablyearlier than in the other Europeanfrontier provinces.
Sisciawas ableto provide Pannoniawith the small changeof the years368-75 in
abundance. 26 Thesuddendecrease 27 is certainlydueto seriouscauses.Thereasons
for the early cessationof monetarycirculation will best be connectedwith the
settlementsoonafter 375 of a very considerablemassof barbariansin Pannonia,
who paralysedeconomic life and the functioning of the administration over
largeareasof the country,andprobablyin thoseareasin particularwhich werethe
most importantlinks betweenthe Danubefrontiers and the South.TheJoederati
in Pannoniawere wedgedbetweennorthernPannonia(PannoniaI and Valeria)
andthe south(Italy andthe Balkans),and this situationmust be significant.
The political activity of the Joederati and their employmentin various elite
units show that their presencein Pannoniaand Dacia by no meansimplied a
peacefulcoexistenceof Roman town-dwellersand of partly barbarianinhabit-
ants of the countryside.Moreover they had chronic difficulties of supply and
343
establishedthemselvesas parasiteson the provinces,which in any case were
economicallynot up to standard.In 391 Theodosius hadto fight against the
Goths in the Balkans;he defeatedthem in 392 nearthe Rhodopemountains.28
It was probablythe Pannonianfoederati who attemptedan attack on north Italy
in the sameyear,29 and this may suggestthat the situation was even worse in
Pannonia,or that Pannoniaitself had little left to offer them. Jerome,observing
eventsin his homelandwith understandableattention,reports now and then a
few details which becauseof the emotional tersenessof his style may appear
mutually contradictory.Historians have gone to either extreme:either every-
thing had alreadyperishedin Pannonia-orratherin the Pannonian-Dalmatian
border regionaroundStridon-by38o, or everythingwas still nice andpeaceful
c. 410. Jeromereportedthat Stridon had beendestroyedby the barbariansin the
year 380;30then under 392 he wrote that 'all perished';3!a few yearslater, how-
ever, he wished to sell his father's estatein Stridon.32 An expressionof pathos
and despairsuchas cunctaperierunt, of course,cannotbe contrastedwith the sale
of 'half dilapidated little farms', as Jerome characterizedhis father's estate.
Semirutaevillulae quae barbarorumeffugerantmanuscan be presumedto be a factual
description of the situation in Pannonia,while generallaments on the damna
Illyrica, the Ilbwica vastitas, etc., have a long-term justification. Otherwisethese
expressionswould not have beenusedso quickly and so widely by contempor-
ary historians. Apart from Jerome,they are encounteredin the works of St
Ambrose, Orosius, Pseudo-Augustineand others,33 and the fact that this
phraseologyoccurslater, too, in SidoniusApollinaris or Ennodius,34does not
affect its truth. The first occasionfor theselamentswas providedby the Gothic
movementsbeforeAdrianople; their repetitionat the end of the fourth century
andin the fifth was causedby new attacksand by continualinternal unrestin the
provincesconcerned.The stagesof declinewere often bewailedin contemporary
literature with conventional expressionsand metaphors,but the underlying
truth must not be doubteddespiteits guiseof literary platitude.The situationin
Pannoniais describedby Claudianin retrospectin 399 as a continualsiege;35 the
inhabitantsof the towns do not dare to leave the town-walls, and the peasant
cannot cultivate his fields. Certainly, this also is an exaggeration,particularly
since Claudian wished thus to emphasizethe merits of Stilicho. But times of
crisis wererepeatedonly too quickly andconsolidationwas hinderedbecausethe
foederati could not adapt themselvesas producersin the economiclife of the
country. Their most importantsourcesof incomewere their positionamongthe
elite troops of the centralarmy, and their exploitationof the provincial popula-
tion, partly by means of the food-rations they drew and partly, no doubt, by
straightforwardplundering.
344
Alatheus and Saphraccontributed important contingents to the army of
Theodosiusin the battle at the Frigidus in 394, in which the Visigoths under
Alaric also took part. The troubles that broke out in the following year, which
were partly connectedwith dissensionsin the camp at the Frigidus, affectednot
only the lower Danubebut probablyPannoniatOO.36 Whetherbarbariansbroke
into Pannoniacannotbe provedfor certain, but Stilicho had to fight with bar-
barianssoonafter he had crossedthe Julian Alps.37 Claudiantells us nothing of
successesby his patron in this connection,but wide-scaleplundering by the
Goths and the first appearanceof the Huns on the lower Danube probably
offered the Pannonianfederates an opportunity to fish in troubled waters.
Archaeologistsin Pannoniahave connectedmany burnt depositsand destruc-
tion layerswith the plunderingsof the year 395; whetherthey are right to do so
cannotbe determined,if only becausethe dating evidenceof coins, as we have
seen,becomesvery rare after 375. However,from Vindobona,where the exist-
enceof thick burnt layers can be established,a comparativelylarge coin-hoard
has come to light of which the final coin datesfrom 395.38 The joederati pre-
sumablydid not limit their raids to the areaassignedto them, but whetherthey
wereresponsiblefor the burningofVindobonacannotbe definitely stated.What
makesit less likely is that the immediateneighboursof Vindobona,the Marco-
manni, are heardof at just this time.
In Paulinus'sbiography of St Ambrose we read of a Marcomannicqueen
namedFritigil39 who, under the influence of the great bishop of Milan, was
converted to Christianity, and persuadedher husband to surrender to the
Romans.It follows from this that part of the Marcomannicrossedto Pannonia,
and was given the receptio,and an entry in the Notitia confirms this settlement
from another angle.40 We read there of a tribunus gentis Marcomannorum in
PannoniaPrima, and it is much more likely that this genswas the peopleof the
husbandof Fritigil than the descendantsof the Marcomannicking, Attalus,
settledin PannoniaunderGallienus.At the sametime this entry seemsto be one
of the last in the Pannonianchapterof the Notitia.
Rome'sold neighbourson the Danubefrontier, the Marcomanni,Quadi and
Sarmatians,and the comparativelynew ones such as Vandals and Gepidaeare
otherwise hardlymentionedat all towardsthe end of the fourth century.For the
most part they only appearin lists of peoplethat had startedto move and were
threateningthe Danubeprovinces,and we cannotthereforedecidewhetherthey
were enteredin theselong lists merely for the sakeof completenessor whether
they had actually taken part in the fighting, 41 as subordinatepartners of the
Goths and Huns. The Sarmatiansare namedfairly frequently by Claudian,42but
generally only to describea geographicalarea or the section of the Danube
345
The beginning~f the Dark Age
known as the ripa Sarmatica.Thesepassagesin Claudianare at the sametime the
only proof that the Court of Ravennaconsideredthe Danubefrontier as one still
held by Romanoccupyingforces.From the Notitia, which is an enigmaticsource
for late Roman military history, we cannot disentanglewhich of the troops
enumeratedin the respectivechapterswere the latest onesand when they were
stationedon the Danube.The lists of troopsin Valeria andPannoniaSecundain
particularseemto be rather old-fashionedin comparisonwith the lists of other
provinces,in that they arein severallayersand containa long successionof fort-
sites enumeratedtwice and three times over.
The problem of the end of the Roman limes, however,is not specialto the
Danubeprovinces.As long as historiansassumeda single 'collapse'of the limes
as the result of a single barbarianattack, a date was sought before which the
limes existedas a statefrontier with regular garrisons,and after which it simply
no longer existedas a frontier and military zone. However,from the end of the
fourth century we must distinguishbetweenformal statefrontiers, frontiers of
actual spheresof influence and strategically important lines occupied by the
army. We hearpractically nothing of the Danubianfrontiers having played any
part in the protection of Roman territory. The territory occupiedby foederati,
andlater by barbarianswho werenot federates,extendedinto former areasof the
empire,but this did not preventthe Danuberemainingthe formal frontier of the
Romanempire. Building periods datableto the fifth century cannotbe proved
on the limes, and the few, technicallyinferior and otherwiseunimportantrepairs
in the forts of the limes which mustbe of post-Valentinianicdatecanonly prove
continueduseof the stonebuildings by variousethnicor socialelementsbut not
regular use of the limes-buildingsby regular Romantroops. The settlementof
foederati on the border gradually deprived the garrisons of the limes of their
strategicimportance; it was less and less necessaryand possible to maintain
discipline, to fill up the units by regularrecruitmentand to secureprovisions.A
gradualdecline,however,can only rarely be followed in detail and divided into
periods.The processof this declinestartedunderTheodosiusandendedin about
the 420S, when the central administrationceasedto attemptto keep the lists of
troops in the Notitia up to date.
In the period that we have beendiscussingthe Danubelimes had no signifi-
cance.It lastedlongeston the border of PannoniaSecundaand Moesia Prima,
where no federateswere settled. But this section of the limes was the least
important on the middle Danube.It gainedimportanceonly when the centreof
the Hunnic empire was shifted to the Danube-Tisza plain, when it was this
sectionof the limeswhich was most exposedto pressurefrom the Huns. But then
it ceasedto be a militarily occupiedfrontier.
346
In 399 or 400 Stilicho is praisedas Rhenipacatoret Histri by Claudian43 and in
the samepoem a number of Stilicho's successesin Pannoniaare listed.44 The
context does not make it certain whether Stilicho visited Pannonia,and the
panegyricexaggerationof his successesdoesnot permit recognitionof the true
state of affairs. According to Claudian the peaceof former times returnedto
Pannonia;but the inhabitantswere not able to enjoy the results of this pacatio
for very long. In 401 the Vandals marchedfrom their homein the north of the
HungarianLowland throughPannoniato the west,45andthis migrationparticu-
larly affected that part of the province which up to that time, though by no
means sparedaltogether,had not been settled by joederati. In the same year
Alaric also departedandmovedwith his peopleto the west. The migrationfrom
Dacia Ripensisthrough Pannoniato Ravennacould be accomplishedby only
one route, throughthe valleys of the Timok, Morava and Save,and then over
the JulianAlps. Alaric arrivedin Italy in 402, andafter he hadto retreatout of it
he was given Pannoniaby Stilicho,46but probablynot as a federateterritory, and
this presumablyaggravatedthe situationof the provincestill further. We do not
know whether Alaric's brother-in-law and successor,Athaulf, had already
settledin Pannoniaat that time as well; but a few yearslater he was certainly in
westernPannonia.47
This secondgreatsettlementof barbariansin Pannonianeitherlastedlong nor
was able to achieveconsolidation.In 405 Radagaisus'huge band of barbarians
movedthroughnorthernPannonia,probably on the route usedby the Vandals
when going west.48 This time it was probablythe Visigothic control of southern
Pannonianearthe approachesto Italy underAlaric and Athaulf which was able
to prevent their crossing the Julian Alps. This migration no doubt had a
catastrophiceffect on the areasof Pannoniawhich lay in its path. A few years
afterwardsJeromecountsthe hostesPannoniiamongthe peoplesthat had poured
into Gaul, adding a sigh over the lugenda res publica.49 The fact that the hostes
Pannoniiare namedat the end of the list, so to speakas a climax, with the added
lament, suggeststhat Jerome meant the non-barbarianinhabitants of Pan-
nonia, though presumablynot the provincial upper class. A few statements
in our sourcesattesta flight of the socially higher-rankinginhabitantsof Illy-
ricum to Italy and presumablyto the Dalmatiancoasttoo. Someof theseare to
be datedafter 403, othersbefore408, and thus are best relatedto the march of
Radagaisus.Thesestatementsare the first to mention Pannonianrefugeeswho
from the beginning of the fifth century onwardsappearedin various placesin
the west and in the Mediterraneanarea.
We can probably recognizeBishop Amantius as one of the earliestfugitives.
He is attestedin 381 as bishop of Jovia, and after officiating for twenty yearsin
347
the Christian community of this small town on the Drave, he returned,some
time in the beginningof the fifth century,to Aquileia, wherehe died in 413. The
inhabitantsof Scarbantiafled from an incursio barbarorum to Italy, and took the
relics of Quirinus,the martyr bishop of Siscia, from Savariawith them. In 402/3
Quirinus was not yet buriedin Rome,50andso we may attributethe paniccaused
by the barbariansin north-westernPannoniato the invasionof Radagaisus.The
fact that the bonesof the martyrswere takenfrom their gravesandaccompanied
the flight is significant enoughfor the situation: the martyrs' graveswere the
focus of cult-life, they were preservedin a Cella Memoriae in the Christian
cemetery(as in Sirmium), or in a basilica(as in Savaria).By the translationof the
relics, the Christian communitieslost their most important treasures.
The flight from Radagaisusseemsto have set in motion a fairly large massof
people. In 408 a decreewas promulgated,which suggestsa large number of
refugeesfrom IllyricumY A numberof fugitives who did not have the appro-
priate proofs of identity fell into an awkwardsituationby being seizedas slaves;
the decreerefers to this situationhaving alreadybecomechronic, and speaksof
the whole of Illyricum, both fugitives and those remaining on the spot. It
follows that in 408 the situation was still unconsolidated,and this conclusion
canalso be drawnfrom other decreeson captivesandrefugees.ProbablyAlaric,
too, took captive Romanswith him to Italy.
About 408 at the latest the SenatorFlavius Lupus was active in Pannonia
Secunda.His duty was to control the land-tax, whose proceedswere used by
imperial decreefor constructionwork on the walls of the towns.52 The towns of
this particular province for the most part presumablyalready possessedtheir
town-walls,like Sirmium, for instance,whosewalls arementionedby Ammianus
under the year 374. But an earlier date for the activity of Lupus in Pannoniais
equally feasible.
What happenedin Pannonia,after the Gothic kings Alaric and Athaulf had
migratedto Italy, is obscure.The only recordwe havefor two decadesfollowing
408 is of a military commandunder a certain Generidus,who, according to
Zosimus,was the commanderof the troops in Dalmatia, Raetia, Noricum and
Upper Pannonia(i.e. PannoniaPrima), about 409, and who proved himself a
brave commander.53 In the Notitia, the ducatesof Noricum Ripense and of
PannoniaPrimaare, in fact, united,and the apparentlyaccurateenumerationof
Generidus'commandin Zosimus may thereforepossibly suggesteither that, of
all the provincesof West Roman Illyricum, it was only in these that troops
remainedany longer,or elseit was only in thesethat Generiduswas in a position
to organizea proper occupation.Possiblyit was under him that Noricum Rip-
enseandPannoniaPrima, which in the Notilia havea commondux, wereunited.
348
The double mentions of forts, which suggestthat the list of troops of the
province in questionis a compilation, hardly occur in the Norican-Pannonian
ducate,and this may meana relatively late validity for this chapter.54
We do not know how long the commandof Generiduscould be maintained.
In the secondand third decadesof the fifth century the areacontrolled by the
empire of the Huns presumablymoved westwards;as a result Pannoniasoon
found itself neighbourto the Huns andprobablyin the third decadeparts of the
province actually cameunder Hunnic rule. About 408 the centreof the Hunnie
empire, of course, still lay on the lower Danube, probably opposite Dacia
Ripensis. But the migration through Pannoniatowards the west by Vandals,
Gothsand the mixture of peoplesunderRadagaisuswas in the last resortcaused
by expansionby the Huns westwards.These movementsof peoplesheralded
Hunnic expansion,and they affectedMoesia Prima which up to that time had
generally been spared.Moesia Prima, Dacia Mediterraneaand Dardania had
often beenexposedto raids by the Goths who had beensettlednorth and south
of the Haemusin 376, but theseraids concentratedmuch more on the areaseast
and south of what had formerly beenUpper Moesia, which was only the peri-
phery of the endangeredarea. Nothing is reported of any emigration worth
mentioningat the beginning of the fifth century, and continuity of the church
hierarchycan thereforebe tracedback in sometowns of Upper Moesiainto the
fifth century,55in contrastto Pannonia,whereexceptat Sirmium56 no bishopis
attestedin the fifth century. The surroundingsof Sirmium sharedthe fate of
Moesiafrom the twentiesof the fifth centuryonwards,and in 437 were cededto
the easternRomanempire,evidentlyfor the simple reasonthat they were cut off
from the westernempire by Hunnic expansionin the areaof Pannonia.
The proof that by 420-5 parts of Pannoniahad come for practical purposes
into Hunnic possessionis given by the chroniclerMarcellinus comes,who under
the year 427 mentions the Roman reconquestof Pannoniawhich had been
occupiedfor the last fifty years by the Huns.57 It is obvious that Marcellinus
countedthe round fifty years'rule by the Huns from 379, that is from the date
when the peoplesof Alatheusand Saphrac,part of whom consistedof a Hunnic
group, obtaineda footing in Pannonia.The Huns who occupiedPannoniain 427
were certainly not these old Hunnie federatesbut the Hunnic empire, whose
centreat that very periodwas movedinto the Hungarianplain. The sameconclu-
sion follows from the fact that two decadeslater the bishop of Margum was
accusedby Attila of having robbed Hunnic princes' gravesin the approaches
of Margum.58 Hunnie princes were thus already buried in the thirties in the
southernparts of the Lowland.
Under suchcircumstanceswe must take the cedingby treaty of large areasof
349
Pannonia59 in 433 to be the legal recognitionof a stateof affairs that had already
existedin reality for years. Furthermore,the cessionof 433 must meanthat in
427 Pannoniawas only temporarily regained.It is very likely that in 427 it was
East not West Rome which regainedparts of the province: Marcellinus comes
composedhis chroniclefrom the view-point of EastRome.When, therefore,in
437 WestRomecededSirmium to EastRome60 this againwasprobablymerelya
legal settlementof a de facto situation. Moreover, since it was possibleto nego-
tiate about Sirmium four yearsafter Pannoniahad beencededto the Huns, the
conclusionis clearthat the Hunnic conquest,sanctionedin 433, did not relateto
the surroundingsof Sirmium, but to Saviaandthe areasnorth of the Drave.We
have alreadyseenthat the federatepeoplesof Alatheusand Saphrac,consisting
partly of Huns, were mainly settled in Savia. The East Roman reconquestof
Pannoniacould be connectedwith a fifty years'rule in Pannoniaby the Huns,
for the reasonthat, in 427, from Sirmium, East Rome undertookan operation
in the areabetweenthe Drave and the Save,that is in Savia.
By c. 437 the largestpart of Pannoniawas in the possessionof the Huns. Pour
yearslater, in 441, Attila conqueredthe EastRomanpossessions on the Danube:
the towns of Ratiaria, Viminacium, Singidunum, Sirmium and Naissuswere
besiegedand taken,and a strip of land five days' journeywide had to be evacu-
ated south of the Danube.61 In 447 the official trading-postbetweenHuns and
East Romanswas moved from the Danubeto Naissus.62
With the stabilization of Hunnic rule in our provinces the history of the
Romanempire on the middle Danubeproperly ends.After the collapseof the
empire of the Huns the Danubelands cameunderthe rule of the Goths who in
their turn were successivelyreplacedby German, Turkish, Slav-andin the
Carpathianbasin alsoby Finno-Ugrian-peoples.Though the EmperorAvitus
did appearin Pannoniain 455, this demonstrationof force hadno results.We do
not even know how far A vitus advancedfrom Italy. For a time south-western
Pannoniabelongedto OstrogothicItaly, and the region of Sirmiumwas retaken
severaltimesby EastRomeuntil it was finally lost at the endof the sixth century.
Prom what Procopiussays63 we may presumethat betweenthe time of Attila
andthat of Justinianthe Moesianprovinceson the Danubesufferedmuch. Only
Dardania remained permanently in East Roman possessionand southern
MoesiagaveByzantiuma numberof distinguishedmenof the old Illyrican kind:
the EmperorsJustin and Justinian came from Thracian country stock in the
southernpart of the territory of Naissus,not far from the northernborder of
Dardania.The interludeof Justinian'sreorganization,however,lies beyondthe
scopeof this book.
350
This conciseaccountof the courseof events,from the time whenthe Goths and
otherfederatesgaineda foothold within the empireafter the deathof Valentin-
ian down to the periodof the Hunnic empire'sgreatestexpansionon the Danube
underAttila, remainsof coursein manyrespectsuncertain.This is due aboveall
to a changein our sources.The rich polemic literature of the end of the fourth
century contains exceedinglyvaluable information on the contemporarysitu-
ation, and in the occasionalpoetry of the court poet Claudian fairly accurate
statementscanalso be found. But thereaftereverythingbecomessilent. For later
times we haveonly short, occasionalnoticesor out-of-contexttraditionswritten
down at a much later period. The situationin Pannoniaduring the early period
of joederati is describedby severalwriters with accuracyand feeling; but of the
advanceof the Huns intoPannonia,hardly a quarterof a centurylater, we have
only chancereports, made by authors whose interestswere focused on quite
different things. In my opinion, this alterationin the interestsof contemporaries
is an indication of Rome'sdecreasinginterestin Pannoniaas a possession.One
has only to think of SidoniusApollinaris praisingthe enterpriseof his father-in-
law A vitus in Pannonia(in 45 5) as the reconquestof a country that Rome had
lost many generationsago.64 For Procopius Pannoniawas already a desert
devoid of people,65and that, though in no way true, is neverthelesssignificant
for a Byzantine'sway of thinking underJustinian.The lamentsover the loss of
Pannonia,repeatedlyto be heard from the time of Theodosius,are not only
characteristicof the attitude of the Romans towards Pannonia, but of the
situation of Pannoniatoo.
As far as the former Upper Moesia is concerned,similar statementsin late
Romanliteratureare neither frequentnor tragically portrayed.A crisis such as
Pannoniaexperiencedin thefirst decadeof the fifth centurydid not befall Upper
Moesia until the time of Attila. Whilst federatesand passingbarbariansplun-
deredPannoniafar andwide, thereis no mentionof oppressionin UpperMoesia
in the works of Nicetas,bishop of Remesianaat that time. The Romanciviliza-
tion of Moesia experiencedits time of terror only after the middle of the fifth
century,andwas thereforesparedto a greaterdegreethanthat ofPannonia.This
difference in phasing provided a firm foundation for the reorganizationof
Justinian,which in all probability could not have succeededin Pannonia.The
Romaninhabitantsof Moesia had stayedon the spot in larger numbers.From
Priscus'report of his journeysomedetailscan be learnt of the towns of northern
Moesia: Naissusat that time was still inhabited,and Ratiaria, accordingto Pris-
cus, was a large and populoustown. Many late antiqueor Christianinscriptions
from Viminacium, Naissusand Ratiaria can be datedto the fifth century,66but
in Pannoniasuch a date can be suggested,if at all, only for a few inscriptions
35 1
from Sirmium. To judge by its fate and its political relationship,Sirmium had
belongedto the easternempirefrom c. 420. Thus continuity of churchhierarchy
is provedin Pannoniaonly at Sirmium, whereasin UpperMoesiait is established
at severaltowns: in 424 the bishopsof RatiariaandViminacium, in 449 a bishop
of Remesiana,and in 458 a bishop of Scupi are attested,to say nothing of the
bishopof Margum who was involved with the Hunnic princes'graves.Someof
theseseeswere extinguishedin the later fifth century,but Justinianendeavoured
to re-establishthem.67
As the fates ofPannoniaand UpperMoesiawere basicallyso entirely different
from after the end of the fourth century, we shall study eachseparately.
The developmentin Pannoniacan be divided into periods approximatelyas
follows (Sirmium for reasonsgiven is not takeninto consideration):(i) 379-401:
settlementof the Goth-Alan-Hunfederatesof Alatheus and Saphrac,leading
to repeatedserious crises and devastations;(ii) 40I -8: Pannoniabecomesthe
corridor of various Gothic and other peoples: emigration and flight of the
Romansbegin; (iii) c. 4°9: a short period of quiet, with attemptsto organizea
Romanoccupation;(iv) from c. 410-20 onwards:gradualexpansionof Hunnic
control over Pannonia,ratified by the cessionby treaty in 433.
The Upper Moesianprovinces,on the other hand, were settledonly in part
with Gothic federates:in Dacia Ripensis,and probablymuch more in what was
formerly Lower Moesia. The periods of developmentin Upper Moesia can
thereforebe put as follows: (i) 376-4°1:settlementsof federatesin parts of the
country only, causingunrestfrom time to time, but life on the whole was much
more peaceful than in Pannonia;(ii) 4°1-10/20:northern Moesia becomesa
corridor for Goths pressingwestwards,but without temporary settlementas
happenedin Pannonia,nor causinglarge-scaleemigration; (iii) 420-41: Upper
Moesiais controlledby East Rome,and in the main is sparedby the Huns; (iv)
441 : conquestby the Huns and evacuationof northernMoesia.
In the light of these differences, the fate of the Romanizedpopulation in
Pannonia and in Upper Moesia becomes easily understandable.From 379
onwards the Pannonianprovinces had had to suffer so much, that the short
period of peaceafter 408 was of only limited advantageto them. The long
coexistenceof barbarianfederatesandRomanprovincials,disturbedby repeated
crises and pictured by contemporariesas a state of siege, was followed by the
heavy ordealsof 401-8 when the country becamean invasion corridor both in
the north and in the south. It is not surprisingthat after all theseordealsmany
Romansfled, thus causinga recessionin secularand ecclesiasticallife alike. The
flight and emigration from Pannonia began after two hard decadesof the
obsidio, and continuedwithout interruptionuntil the invasionof the A vars. The
35 2
fact that some Romansdid not emigrateuntil the later fifth or even the sixth
centurydoesnot by any meansimply that all Romansheld out till the appearance
of the A vars; equally unjustifiablewould be the inferencethat the emigrationin
the first decadeof the fifth century meansthat thereafterno Romanswere left
in Pannonia.The fifth and sixth centurieswere a period of gradualemigration;
I am inclined to assumethat the largestemigrationwas the earliest.Nor mustwe
forget that the descendantsof the followers of Alatheus and Saphracstayedin
Pannonia,and that the comparativequiet after 408 was enjoyed by a Roman
populationthat was alreadydecimatedby emigration.
Both emigrationin groupsand sporadicindividual flight canbe deducedfrom
various indicationsin the sources.We have alreadymentionedthe flight of the
Romansof the town of Scarbantia,taking the relics of Quirinus with them to
Rome.68 In date it presumablycoincideswith the flight of large sectionsof the
population before Radagaisus;an edict of December408 deals with these
refugees.69 Bishop Amantius had possiblyfled evenearlier, c. 400, to Aquileia.70
Just as Quirinus found a final resting-placeon the via Appia at Rome, further
Pannonianmartyrs came to Rome and to severalother towns of Italy.7I They
were brought by the refugees,and this in effect signified the dissolutionof the
Christiancommunitiesin Pannonia.Continueduseof Christiancult-buildingsin
Pannoniaafter the fourth century doesnot prove thesurvival of the Pannonian
cult-communities.Goths and other peoples were Arian Christians, too, and
continuedto use the dilapidatedbasilicas,thoughso far this can only be proved
at the so-calledBasilicaII in the largefortified estate-centreat Fenekpuszta. 72 At
the majority of sites it is not possible to prove unequivocally whether the

reconstructionof a basilica is to be attributed to barbariansor to Romans
stayingbehind.As far as Fenekpusztais concernedthe recentand older excava-
tions have brought to light a unique group of finds that suggestnot only
Christiansbut Romanstoo. This so-calledKeszthelycultureis a specialarchaeo-
logical assemblageof the early A var period which cannot be satisfactorily
tracedback to the local culture of the Romanperiod. Thereare, however,finds
such as a gold pin with the name BONOSA (PI. 45 b), proving that some ethnic
group of Roman complexion remainedat Fenekpuszta. 73 An exampleof such
sporadicRomanswho had stayedbehindin the fifth centurywas St Antonius of

Lirina, who was born in the province of Valeria towardsthe middle of the fifth
centuryand as an orphanedchild was brought to his uncle Constantius,bishop
of Lauriacum. The latter must therefore have already emigratedto Noricum
earlier.7 4 Even later Leonianustravelled from Savariato Gaul.7S St Martin of
Bracarain Spainis also considereda Pannonian. 76 Even the Lombardsdragged
Romansalong with them to Italy, if credit can be attachedto the statementby

353
Paulus.77 The last emigrationwas that of the Sirmiansto Salona,as mentioned
there on someinscriptions of the beginningof the seventhcentury.78
In view of the early beginningof barbarizationand the gradualemigrationof
the Romans,it becomesunderstandable that the place-namesof Pannoniawere
preservedto the presentday only in exceptionalcases(Fig. 58). The namesof
rivers, of course,are the same. (Danuvius = Danube,Donau, in Hungarian
Duna; Dravus = Drave, in Hungarianand south Slavonic Drava; Savus =
GermanSave,in HungarianSzava,south Slavonic Sava; Sala = in Hungarian
Zala; Arabo = GermanRaab,in HungarianRaba; and in this group must be
included the Hungarianriver-nameMarcal which goesback to Mursella. Even
the namesof rivers beyondthe frontier are the old ones:for instanceGranua =
GermanGran, in HungarianGaram,in SlovakianHron). Romanplace-names,
however, which have survived to the present time are representedonly in
Ptuj = Pettau(Poetovio),Sisak(Siscia),Raab(Arrabona,thoughin Hungarian
this town is called Gyor) and possiblyWien, Vienna (Vindobona)althoughthis
is very questionable.All these placeslie on the edge of Pannonia.Not even
Sirmium hasbeenpreservedas a town-name,althoughthe areabetweenDanube
and Saveis namedSrijem (in GermanSyrmien,in HungarianSzoreny,Szerem-
seg) from it; the town was later namedafter its great martyr: in Hungarian
Szavaszentdemeter ('St Demeter on the Save'), in south Slavonic Mitrovica
(Dimitrovica).
The situationin UpperMoesiais a little different: thereare two areastherein
which the ancientplace-nameshavebeenwell preserved,and not only thoseof
the largestplaces.One of the areasis part of Dacia Ripensis,whereArear goes
back to Ratiaria, Vidin to Bononia,Flortin to Florentiana,Rgotina to Argent-
ares. The other areais the southernand easternpart of DardaniawhereNaissus
is nowadaysnamedNis, Scupi Skoplje (in Turkish Uskub) and a place not far
from Ulpianum Lipljan. The place-namesof Moesia were all still in use in the
sixth century; their disappearance thereforeis to be attributedto the upheavals
causedby the Slav conquest.The survival of namesin the caseof large towns,
which alwayshadan importanceas junctions(Naissus,Scupi,Ulpianum), canbe
comparedwith the survival of the namesof Poetovio,Siscia and Sirmium. But
for the survival of the small group of place-namesin Dacia Ripensistheremust
have been a special reason.In particular we may think of the settlementof
Romanizedelementsin the new Dacia, after the evacuationof the Dacia of
Trajan,wherebya Latin-speakingpopulationwas created,which by comparison
with the otherDanubianand Balkan countrieswas compact,if not massive,and
could resistthe Thracianlanguageof the native populationand the influenceof
the Germanic,Bulgarian-Turkishand Slav population.Ratiaria,moreover,was
354
+ Town; Settlement o Drava < Dravus

KO Kulpa < Colapis
P KR krka < Corcoras
M MarcaJ< Mursella
:R.
+ P Ptuj < Poetovio
.. ;.;, R Raab,Raba< Arrabo(na)
-i:...
+ S Sisal< < Siscia
SA Sava < SaVU5
P +SR Srem < Sirmium
P V Vuka. < Hiulca
P
1" W Wien ? Vindobona
P P P Z Zala < Sala
P
P
+A Arter < Ratiaria

P
+F Flortin < Florentiana P
+l Lipljan < Ulpiana P
P P
+N Nis.Ni!ava < Naissus P
p Pek < Pincus P
+R Rgotina<Argentares
+5 Skoplje < Scupi
T 7imok < Timacus
+v ViC/in < B'din < Bononia P
P
+S
0 100
, 200
, 34° km'
Figure 58 Geographicalnamesof ancientorigin
an old military colonia where the Latin languagealways had had strongerroots
than in the municipia with their preponderantlylocal inhabitants.
That theseRomanswere the precursorsof the medievalVlachsis known; but
the subjectis beyondthe scopeof this work. The Vlachs, a Balkanpeople,who
createda neo-Latinlanguageand who are today known as Vlach groupsin the
355
Balkan peninsula,and, of course,as Roumaniansin Roumania,originatedfrom
a people of Latin-speakingmountain-pastoralists;possibly in the case of late
Roman Moesia we can identify their origin. When Scupi was affiicted by an
earthquakein 518, the inhabitantsof the town stayedin the mountains,whither
they hadfled from the barbarians. 79 In this way the Romanizedpopulationof the
town could graduallyadoptthe age-oldDardanianway of life that was basedon

hill-pastures.The fact that lowland settlementshad to be given up in the course
of the fifth century,andin later times evenmore so, andhadto be built afreshon
mountainsand hill-tops for reasonsof safety,is provedby the building activity
of Justinian.80 The constructionof fortified settlementsunderhim took the new
developmentinto accountto a large extent, but it was for this very reasonthat
the Justinianic reorganizationcan only have been an interlude. What had
perishedin the fifth centurycould not be reconstructed,andthe survivorsof the
Roman population necessarily started a new development, dictated by the
changedsituation. Thoughthis enabledRomanelementsto survive till modern
times in the shapeof the Balkan Vlachs and in the Roumanianswho derived
from them, it could not provide the Justinianic reorganizationwith a viable
basis.
In Pannoniageographicalconditionscould not haveprovidedthe Romanized
populationwith a similar possibility of continuedexistenceby changeto upland
pastoralism,evenif that populationhad remainedas a whole in their homeland.
But the emigrationof the Romansfrom Pannoniahas deeperreasonswhich lie
in the specialdevelopmentof Pannonia.More than once we have beenable to
show that the upperclassof provincial societyin Pannonia,which was the only
Romanizedgroup there, enjoyed an existencewhich was largely tied to the
interests of the imperial government.Moreover, the leading class of society
owed its position to its possessionof a sourceof income outsidethe province,
whether as membersof the army or as representativesof some branch of the
administration.Thus, its basisof existencesurvived only as long as its connec-
tions with the central governmentwere not put in jeopardy. As soon as the
foederati intervenedas a new element,and connectionswith the central govern-
ment loosenedgradually,this social classlost its basisof existence.That was the
reason why the Romanizedclassesof Pannoniatook to flight or emigrated
gradually.An important detail must be added.We havealreadypointedout the
likelihood that a largepart of the greatestatesof Pannoniawere imperial estates.
As a result,therewasno numerouslandednobility in Pannoniato remainon the
spot,like the senatorialaristocracyof Gaul or Spain,to becomean equalpartner
with the barbariantribal aristocracy.Thus the local Romanizedupperclasstook
less and less sharein the eventsof the empire'shistory and in that of their own
35 6
country; the active role was taken by the barbarians.This, too, may have been
the reasonwhy Rome had to relinquish Pannoniaso early.
The situationas just describedis presumablyapplicableto Moesia,too. There,
however,the membersof the leadingclasshad hadno reasonat first for emigra-
tion, and when the Huns obtainedthe upper hand they found a new way of
preservingtheir old leading role. This way was also availableto the Romansof
Pannoniawho had stayedbehind,down to the time of Attila, andin this respect
there are no differences between Pannonianand Moesian Romans. For the
Illyrican Romansalso this was the last chanceto take their sharein the direction
of events,and thus to preservetheir leading role. It is well known that Pan-
noniansand Moesiansassembledat the court of Attila in great numbers.The
last emperorof West Rome,too, camefrom a circle which, originating as it did
from Sirmium, took sideswith Attila. 8r Priscustells of the trader from Vimina-
cium who gives a charminglittle accountof his reasonsfor going over to the
Hunnic camp.82Towards the middle of the fifth century the Romansof Illy-
ricum had no choice but to place themselvesat the disposalof the Huns, who
were in need of specialistsand who probably paid well too. This was brain-
stealingwhich in the last resort deprived Rome of capableminds, but for the
RomanizedclassesunderHunnic rule it was the only possibility of arrestingtheir
declineanddissolutionin the broadun-Romanizedmassof people.It was a short-
lived expedient.After the collapseof the Hunnic empire theseRomanshad to
flee. It was probably thus that St Severinuscameto Noricum,83 where he was
able to exert his very extensiveconnectionsin the interestsof the Romans,but
had to keep his past in the camp of Attila secret. The disintegrationof the
Hunnic empirethus contributedto the declineof the Pannonianas well as of the
MoesianRomans.
In the fifth centuryRomanizedelementsprobablyconcentratedmoreandmore
behind the walls of larger towns. Possibly here too the cohesiveforce of the
commonlanguageoperated.It is an observationof the utmost significancethat
it is not until the sixth century that barbarianfinds can be demonstratedin
Pannoniantowns.84 It follows from this, of course,that the remnantsof the
Roman population couldmaintain themselvesin the towns of Pannoniaonly
during the fifth century. An equally important symptom is the practically
completeabsenceof finds of Roman characterof fifth-century date; all finds
datableto the fifth centuryare of distinctly barbariancharacter.This absenceis
the result of gradualpauperizationand of the decline of industrial production.
Productionwas insignificantin quantity, and the resultswere twofold: finds are
small in quantity and old forms subsist. Owing to the impossibility of distin-
guishing betweenRomanfinds of the fourth and fifth centuriesit is difficult to
357
prove the decreaseand pauperizationof the Romanizedpopulationby archaeo-
logical means.Nor do imports exist from the south and west.
Accordingly one might expect that the local native populationwere able to
resume the use of their Celtic, Illyrian and Thracian languagesin the fifth
century,for the tenacioussurvival of the local languagesof the original popula-
tion can be proved in Pannoniaas well as in Moesia. Jeromewas still able to
speakthe Illyrian dialect of his homeland(gentilis sermo).85Illyrian nomenclature
can be provedwell into the Middle Ages,86andin the monasteriesof the Orient
a Thracianliturgy in the lingua Bessicawas introduced.87 Even in the later fifth
century the inhabitants of the Dardanian countryside still carried Thracian
names;the two travel-companionsof the future emperorJustin were named
Zimarchosand Ditybistos88-genuinely Thraciannamesthat were presumably
in useat the time of the Slav annexationsin the Balkanpeninsula.But from the
endof thefourth centurythe rural populationbecamemoreandmoreintermixed
with barbariansand they lacked the political and social basis for conscious
preservationof their native tongue. For the provincial upper class the Latin
languageformed not only the meansandprerequisiteof their socialpositionbut
also the symbol of social relationshipand at the sametime a force of cohesion.
Justas the Sarmatiansof the HungarianLowlands,who arestill identifiablein the
second half of the fifth century,89 later vanished without trace among the
Gepidaeand A vars, so the Celts, Illyrians and Thraciansdissolvedin the seaof
later conquerors,simply for the reasonthat during the long period of Roman
rule they had lost their native culture and were unableto utilize their language
as a meansto a political life of their own.
358
Ab breviations
Acta Ant. Acta Antiqua AcademiaeScientiarumHungaricae (Budapest).

Acta Arch. Acta ArchaeologicaAcademiaeScientiarumHungaricae (Budapest).
AI] V. Hoffiller and B. Saria, Antike Inschriften aus ]ugoslawien (Zagreb,
1938).
Ann. Ep. L' AnneeEpigraphique(Paris).
Ant. Tan. Antik Tanulmd'!Jok-StudiaAntiqua (Budapest).
Archiiol. Anzeiger ArchiiologischerAnzeiger,see]ahrb. DAI.
Arch. Ert. ArchaeologiaiErtesito" (Budapest).
Arch. Iugosl. ArchaeologiaIugoslavica (Belgrade).
Arch. RoZhl. ArcheologickeRozhle4J(Prague). •
ARD see Ferri.
Arh. Pregl. Arheolofki Pregled(Belgrade).
Arh. Vestn. Arheolofki Vestnik-ActaArchaeologica(Ljubljana).
Barb-Festschr. Festschriftfiir A. A. Barb (Eisenstadt,1966).
Barkoczi, Brigetio L. Barkoczi, Brigetio (Diss. Pann. ii, 22, Budapest,1944-51).
Beitrage see Patsch.
Bevolk. A. Mocsy, Die Beviilkerungvon Pannonienbis zu den Markomannenkriegen
(Budapest,1959).
Bp. Miieml. M. Horler (ed.), BudapestMiiemtekei, Vol. ii (Budapest,1962).
Bp. Reg. BudapestRegisegei(Budapest).
Bp. Tikt. A. Alf6ldi, Gy. Laszlo, L. Nagy, T. Nagy, J. Szilagyi and F. Tompa,
BudapestTo"rifnete, Vol. i (ed. K. Szendy,Budapest,1942).
BRGK Bericht der Ro"misch-Germanischen Komission des DeutschenArchiiologischen
Instituts (Frankfurt).
BronzegeJasse see Radnoti.
359
Abbreviations
Burgenl. Heimatb. BurgenliindischeHeimatbliitter (Eisenstadt).
C. CIL iii (Berlin, 1873-19°2).
CAH The CambridgeAncientHistory (Cambridge).
Carn.-Jb. Carnuntum-Jahrbuch(Graz-Koln).
Carnuntum see Swoboda.
Chron. Min. Chronica Minora saec. iv, v, vi, vii, ed. Th. Mommsen, Berlin, I892-8,
Vols i-iii (Monumenta Germaniae Historica, Auctorum Antiquissimorum
tom. ix, xi, xiii).
CIL Corpus Inscriptionum Latinarum (Berlin).
Corpus Scr. Eccl. Lat. Corpus Scriptorum EcclesiasticorumLatinorum
DCU see Dobias.
Diss. Pann. Dissertatjones Pannonicae ex Instituto Numismatico et Archaeologico
Universitatis de Petro Pdzmdtry nominatae Budapestinensisprovenientes
(Budapest).
Dittenberger,Syll. W. Dittenberger,SyllogeInscriptionum Graecarum,2nd ed., Berlin.
Dobias, DCU J. Dobias,Dijiny ceskoslovenskeho uzemipredvystoupenimSlovanu(Prague,
1964).
Dob6, Verwaltung A. Dob6, Die Verwaltung der riimischen Provinz Pannonienvon Augustus
bis Diocletjanus(Budapest-Amsterdam, 1968).
Egger, RAFC R. Egger, RiimischeAntikeundfriihes Christentum.AusgewiihlteSchriften,
Vols i-ii (Klagenfurt, I962-3).
Ferri, ARD S. Ferri, Arte Romanasui Danubio (Milan, 1933).
Fol. Arch. Folia Archaeologica(Budapest).
Garasanin,Nalazi1ta D. and M. Garasanin,Arheol01kanalazi1ta u Srbiji (Belgrade, 1951).
GMKM Godi1'!Jak Muzeja Kosova i Metohija (Pristina).
Grbic, Plastike M. Grbic, Odabrana grlka i rimska plastika u Narodnom MuZeji (Bel-
grade, 1960).
GSND Glasnik SkopskogNaucnogDru1tva (Skopje).
GZM Glasnik Zema!jskogMuzeja, Arheologija (Sarajevo).
IBAD Izvestija na }3algarskoto ArheologiceskoDruzestvo-Bulletinde la Societe
ArchfologiqueBulgare (Sofia).
IBAI Izvestija na BJlgarsktja Arheologiceski Institut-Bulletitl de l'Institut
ArchfologiqueBulgare (Sofia).
ILCV E. Diehl, Inscriptiotles Christianae Veteresi-iii (Berlin).
ILJug. A. and J. Sasel,'InscriptionesLatinae quaein Iugoslaviainter annos
MCMXL et MCMLX repertaeet editae sunt', Situla 5 (Ljubljana,
1963)'
ILS H. Dessau,InscriptionesLatinae Selectaei-iii (Berlin).
Inscr. teg. see Szilagyi.
Intercisa M. R. AlfOldi, L. Bark6czi, G. Erdelyi, F. Fiilep, A. Radn6ti,K. Sagi,
etc., Intercisa, Geschichteder Stadtzur Rijmerzeiti-ii (Budapest,1954-7).
jahrb. DAI jahrbuch des DeutschetlArchaologischetlItlstituts (Berlin).
JOAI Jahreshe.ftedes OsterreichischetlArchiiologischetlInstituts (Vienna).
jOAIB JOAI, Beiblatt.
JRS Journal of RomanStudies(London).
360
Abbreviations
Kanitz, RS F. Kanitz, Rb'mische Studien in Serbien. Denkschriften der Philos.-hist.
Klasseder Ak.ademiexvii (Vienna, 189z).
Kubitschek, W. Kubitschek, Romerfundevon Eisenstadt(Vienna, 19z6).
Riimerfunde
Laur. Aqu. Laureae Aquincensesmemoriae V. Kuzsinszlry dicatae i-ii (Diss. Pann.
ii, 10-II) (Budapest1938-41).
Limeskongress1969 Congressof RomanFrontier Studies. Cardiff I969 (ed. E. Birley, Cardiff,
1974)·
LJP
LRKN seeVarady.
LRKN Limes RomanusKonferenzNitra I9J7 (Bratislava, 1959).
LuJ Limesu Jugoslaviji (Belgrade).
MFME A Mora Ferenc MuzeumEvkiitryve (Szeged).
Mirkovic, RGD M. Mirkovic, Rimskigradovi na Dunavu (Belgrade,1968).
Monum. Germ. Hist. MonumentaGermaniaeHistorica.' AuctoresAnliquissimi.
Auct. Ant.
MPL Patrologiae Latinae CursusCompletus(Paris) (Migne's Patrologia).
MS A. Mocsy, Gesellschaftund Romanisationin der rbmischenProvinz Moesia
Superior (Budapest-Amsterdam, 1970).
Nagy, Mumienbegrab- L. Nagy, Mumienbegriibnisseaus Aquincum(Diss. Pann. i, 4, Budapest,
msse 1935)·
Nagy, PS L. Nagy, Pannonia Sacra, in SzentIstvdn Emlikkiitryv, vol. i (Budapest,
1938).
Nalazilta see Garasanin.
Not.Occ. Notitia dignitatum ... in partibus Occidentis(ed. O. Seeck).
Not. Or. Notilia dignitatum ... in partibus Orientis (ed. O. Seeck).
Num. KiiZI. NumizmatikaiKbZlOtry (Budapest).
OMRTE AZ OrszdgosMaf!)ar RegeszetiTdrsulat EVkb'nyve-Jahrbuchdes Ungari-
schenArchiiologischenGesellschaft(Budapest).
Osj. Zb. OsjeCki Zbornik (Osijek).
Pam. Arch. PamdtkyArcheologicke(Prague).
Pannonia A. Mocsy, Pannonia, in PWRESuppl., Vol. ix (Stuttgart, 196z).
Patsch,Beitriige C. Patsch,'Beitragezur Volkerkunde von Siidosteuropa'in Sitzungs-
berichte der Akademie,Phil.-hist. KI. (Vienna, 19z8-37).
P. Dura C. B. Welles, R. O. Fink, J. F. Gilliam, The Excavationsat Dura-
Europos,Final Report V, Part i: The ParchmentsandPapyri (New Haven,
1959)'
Plastika see Grbic.
PS see Nagy.
PWRE Pau!ysRealen~yclopadie neubegonnen
der classischenAltertumswissenschaft,
von G. Wissowa,fortgeflihrt von H. Kroll, K. Ziegler, etc. (Stuttgart).
Radnoti, Bronzegefiisse A. Radnoti, Die rbmischen Bronzegefiissevon Pannonien (Diss. Pann.
ii, 6, Budapest,1938).
Radvqjv. muzo Rad vqjvocijanskihmuzda(Novi Sad).
RAFC see Egger.
Regesten see Seeck.
361
Abbreviations
RGD see Mirkovic.
RIC H. Mattingly, E. A. Sydenhamand others, RomanImperial Coinage
(London).
Riv. Ital. di Rivista Italiana di Numismatica.
N umismatica
RLiG Der RomischeLimesin Gsterreich (Vienna).
RomerJunde see Kubitschek.
RS see Kanitz.
RVP see Thomas.
Saxer, Vexillationen R. Saxer, 'Untersuchungenzu den Vexillationen des romischen
Kaiserheeres',EpigraphischeStudieni (Cologne-Graz,1967).
Schober A. Schober,Die romischenGrabsteinevon Noricum und Pannonien(Vienna,
192 3).
SCIV Studii si cercetari de istorie veche (Bucharest).
Seeck,Regesten O. Seeck,Regestender Kaiser und Papste(Stuttgart, 1919).
SHA ScriptoresHistoriae Augustae.
Slov. Arch. SlovenskaArcheolrfgia (Bratislava).
Spomenik Srpska(kraljevska) Akademija,Spomenik(Belgrade).
Stud. Mil. Studienzu den Militargrenzen Roms. Vortrage des 6. InternationalenLimes-
kongressesin Siiddeutschland(Cologne-Graz,1967).
Swoboda,Carnuntum E. Swoboda,Carnuntum.Seine Geschichteund seine Denkmaler(4th ed.,
Graz-Cologne,1964).
Syll. see Dittenberger.
SymposiumI!)64 Simpozijum0 teritorijalnom i hronolofkomrazgranice,yuilira u praistorijsko
doba-Symposium sur la delimitation territoriale et chronologiquedesIltJriens
al'epoqueprehistorique,IJ-I6 mai, I!)64 (ed. A. Benac, Sarajevo,1964).
SymposiumI!)66 Simpozijum0 ilirima u anticko doba-Symposium sur les IltJriens al'epoque
antique, IO-I2 mai, I966 (ed. A. Benac,Sarajevo,1967).
Szilagyi, Inser. teg. J. Szilagyi, Inseriptiones tegularum Pannonicarum (Diss. Pa1tn. ii, 1,
Budapest,1933).
Thomas,RVP E. B. Thomas,RomischeVillen in Pannonien(Budapest,1964).
Varady, LJP L. Varady, Das letzte Jahrhundert Pannoniens 376-476 (Budapest-
Amsterdam,1969).
Verwaltung see Doba.
Vexillationen see Saxer.
VHAD Vjesnik hrvatskogaarheolofkogadruftva (Zagreb).
VMMK A VesZpremMegyei MuzeumokKozlemitryei(Veszprem).
ZA Ziva Antika-AntiquiteVivante (Skopje).
ZFF Zbornik FilosofskogFakulteta Universitetau Beogradu(Belgrade).
ZNMB Zbornik RadovaNarodnogMuzeja (Belgrade).
362
Notes
CHAPTER I
I Homer, Iliad xiii, 4-6.

2 Strabovii, 3, 2-10.
3 S. Borzsik, 'Die Kenntnissedes Altertums tiber das Karpatenbecken',Diss. Pann. I, 6
(Budapest,1936).
4 Typical is, e.g., Apollonius Rhodius, Argonauticaiv, 325-6. Further passagesin Borzsik,
0p. cit., 16-18.
5 See note 80.
6 Hecataeus,fragment 92 (Jacoby). HerodotusV, 9. Apollonius Rhodius, Argonautica iv,
320 , 324·
7 M. V. Garasanin,ZFF vii (i), 1963,45,tracesthe name Singidunumback to the Sindoi.
8 J. Harmatta,Arch. Ert. xciv, 1967, 133-6.
9 J. Gy. Szilagyi, 'Trouvailles grecquessur Ie territoire de la Hongrie: Le Rayonnementdes
civilisations grecque et romaine sur les cultures peripheriques', Vnle Congr. Internat.
d'Archiol. Class. Paris 190J, 389-90; Ant. Tan. xiii, 1966, 102-6. On Greek finds in the north-
ern Balkans seeM. Parovic-Pdikan,Arch. Iugosl. v, 1964, 61-81.
10 iv, 49.
1I Strabo vii, 5, 11.
12 Polyaenusvii, 42, and Theopompusin Athenaeusx, p. 443 A-C, where, however, the
nameof the Autariataewas wrongly handeddown and then incorrectly emendedto Ardiaioi.
A. M ocsy, Rivista 5 torica dell' Antichitd ii, 1972, 13 ff.
13 Strabovii, 3, 8. Justin xxiv, 4.
14 Details in Chapter3 (pp. 58 ff., 63 ff.).
15 Seein particularthe works of J. Untermann,H. Kronasserand R. KatiCic listed in biblio-
graphyby G. Alfoldy, Die Personennamen in der rom. Provinz Dalmatia (Heidelberg,1969), 383-5.
16 M. Suic, RadJugoslovenskeAkadem!ieznanostii umjetnosticccvi, 1955, 121-81.
363
Notes to pp. 4-9
17 G. Alfoldy, Beitriige zur Namenforschungxv, 1964, 55 if., but chiefly R. KatiCic in several
essaysand recently Godifryak iii (Centarza BalkanoloskaIspitivanja, I, Sarajevo,1965), 54-76.
18 Z. Marie in severalworks; seeespeciallySymposium[g04, 191-213.
19 Strabovii, 3, 13.
20 Strabovii,3,10.In general,I. I. Russu,Die Spracheder Thrako-Daker(Bucharest,1969),40f.
21 This theorywas put forward by V. Georgiev.See,e.g., his Trakijskijat ezik (Sofia, 1957).
A map of toponymoussuffixes appearsin Arheofogija (Sofia), ii, 1960, Part 2, 13 if. See also
La Top01!Jmieanciennede fa peninsufeBafkaniqueet fa thesemediterraneenne(Sofia, 1961). Cf. also
note 41 and A. Fol, Rivista Storica dell'Antichitd i, 1971, 3 if.
22 Illyrians in Strabovii, 5, 6.
23 For instance,Dassius,Luccaius, C. 15134; Dasmenus,C. 10212; Terso, Precio, C. 3400.
24 Strabovii, 3, 2; 5,2; 5,10.Seealso Florus i, 39, RufusFestusbrev. 9, AmmianusMarcellinus
xxvii, 4, 10.
25 A. Mocsy, Symposium[g66, 195-200.
26 Livy v, H.
27 Justin xxiv, 4.
28 See map of finds in I. Hunyady, 'Die Kelten im Karpatenbecken',Diss. Pann. ii, 18
(Budapest,1944); cf. T. Nagy, Acta Arch. ix, 1958, 350 if.
29 Strabovii, 3, 8.
30 A. Mocsy, Acta Ant. xiv, 1966,90-2.
31 Theophrastusin Seneca,quaest.nat. iii, II, 2.
32 xxiv, 6, 1.
33 xxii, 9, 1.
34 Justin xxiv, 6; Pausaniasx, 19, 7.
35 Justin xxv, I, 2; Polybius iv, 46, 1.
36 J. Gy. Szilagyi, op. cit. (note 9), 389.
37 Mocsy, op. cit. (note 30), 93 f.
38 Justin xxxii, 3; Athenaeusvi, p. 234.
39 For the Celtic finds in the Save valley, see Z. Marie, GZM xviii, 1963, 83; for those in
Scordiscanterritory, seeJ. Todorovic, Kefti ujugoistocnomEvropi (Belgrade, 1968).
40 Justin xxxii, 3; Diodorus xxii, 9,3; 18; Athenaeusvi, p. 234; Pausaniasx, 23.
41 If there is any justification for dividing the Thracian ethnic group, then, unlike V.
Georgievwho suggestssplitting it into the Thraco-Getaeand the Daco-Mysi (seenote 21), I
considera division into the Thraco-Mysi and the Daco-Getae themore likely. In antiquity
the Dacians and Getae were taken to be one and the samerace in the main, and the native
population of Moesia Superior seemsto have been the western Thracian element. In later
imperial times the native populationin Scupi still spoke the lingua Bessica.SeeMS 249 f. The
personalnamesof the Upper Moesiannativesare Thracian,without any notablelocal colour-
ing: A. Mocsy, 'Vorarbeiten zu einem Onomasticonvon Moesia Superior', Godifryak iii,
Centarza BalkanoloskaIspitivanja 6 (Sarajevo,1970), 175.
42 The ancientswere alreadywell awareof theseand other identical names:Strabo vii, 3, 2;
xiii, I, 21.
43 Justin viii, 6, 3.
44 Livy xxxviii, 16.
45 Livy xl, 57,6; Justin xxx, 4, 12; Pliny, Nat. Hist. iv, 81.
364
Notes to pp. 9-r8
46 For incidents which have beenhandeddown, seeA. M6csy, Acta Ant. xiv, 1966,94-6.
47 Livy xl, 57-8.
48 Livy, periocha xliii and xliii, 18.
49 Seenote 23.
50 Dittenberger,Sylloge 710.
5I Livy xlv, 29, 12.
52 Julius Obsequens16.
53 Livy, per. Oxyr. 174-5; Julius Obsequens22.
54 Pannonia, 527-8.
55 Polybius, fragment 64 (122).
56 A. M6csy, op. cit. (note 46), 97-8.
5 7 vii, 5, I 2.
58 Ibid.
59 Appian, Illyric. 22; cf. Pannonia, 528.
60 Strabovii, 2, 2; cf. 3, 2.
61 vii, 5,2; cf. iv, 6, 10.
62 Seenote 55.
63 Strabovii, 5, 3; Appian, Illyric. 14, 22 and passim.
64 R. KatiCic, Symposium1964, 3 I ff.
65 Strabov, I, 6.
66 E.g. Tacitus, Germ. 2S, cf. 42; Strabovii, I, 3; Velleius Pat. ii, 109, 5.
67 Pannonia, 529; B. Benadik, Slov. Arch. x, 1962, 394 ff.; J. Meduna, Pam. Arch. Iii, 1961,
280 f.; cf.liii, 1962, 135 f. From anthropologicalangle, M. Stloukal, Pam. Arch.liii, 1962, 173.
68 E. Kolnikova, Slov. Arch. xii, 1964, 402. ff., seesa connectionbetweenan aesgrave found
at Zabor in Slovakia and the migration of the Boians fromItaly. Whethersheis right has still
to be investigated.Cf. a similar find from south Pannoniain M. Bahrfeld, Der Miinzfund von
Mazin (Berlin, 1902).
69 Cf. B. Benadik, Slov. Arch. xi, 1963, 372 ff.; Germaniaxliii, 1965,63If.
70 Julius Obsequens48.
71 Livy, periocha lxxxiii; Plutarch, Sulla 23, I; Appian, Mithr. 55; Granius Licinianus 35;
Eutropius v, 7, I; Ps.-AureliusVictor, vir. ill. 75, 7.
72 Appian, Ilfyric. 5; cf. A. M6csy, Acta Ant. xiv, 1966, 98, n. 109.
73 Plutarch, Pomp. 41.
74 Strabovii,3,I1;5,2.·
75 Bell. Gall. i, 5, 4·
76 P. Petru, 'HisasteZare Latobikov', Situla xi (Ljubljana, 1971).
77 Bell. Gall. i, 5, but cf. Swoboda,Carnuntum,231.
78 Livy,periochaxci; Florus i, 39; Rufus Festus,brev. 9; Eutropius vi, 2; Orosius v, 23.
79 Frontinus,strat. iv, I, 43.
So Sallust,hist. ii, 80; Livy,periochaxcii, xcv, xcvii; Julius Obsequens59; AmmianusMarcel-
linus xxix, 5,2; Rufus Festus,brev. 7; Eutropiusvi, 2.. 7-8; Eusebius,chron. 152.k (Helm).
SI Rufus Festus,brev. 7; Eutropius vi, 2.
82 Rufus Festus,brev. 7; Jordanes,Rom. 2.16; Sallust, hist. iv, IS; cf. A. M6csy, Acta Ant.
xiv, 1966, 99.
83 Appian, Mithr. 102., 109, 110, 119; Plutarch, Pomp. 41.
365
Notes to pp. I8-24
84 Strabovii, 3, I I ; Suetonius,Caes. 44; Appian, be!!. eiv. ii, 110.
85 Appian, bell. eiv. iii, 25; cf. ii, IIO, and for Caesar'smilitary plans Strabo vii, 3, 5; Livy,
periocha cxvii; Suetonius,Caes. 44; Aug. 8.
86 Dio xxxviii, 10; Cicero, pro Sestio43 (94); Julius Obsequens61.
87 vii, 3, 11.
88 Be!l. Gall. vi, 25; cf. Pannonia, 532.
89 iii, 8, 3.
90 Pannonia, 532-3.
91 vii, 5, 2.
92 Strabovii, 3, II; 5, 2. Pliny, Nat. Hist. iii, 147. On the problemof this term, seeB. Saria,
in Omagiului C. Daieovieiu (Bucharest,1960),497,n. 2, and Mocsy, Pannonia, 533.
93 On the distribution of Dacian pottery in the Carpathianbasin (outside Dacia), see A.
ToCik, Arch. RoZhl. xi, 1959, 843 fr.; M. Lamiova-Schmiedlova,Slov. Arch. xvii, 1969,459fr.;
S. Nagy, Radvo/v. muzoix, 1960, I I 2 f.; B. Benadik,Germaniaxliii, 1965, 79 f.; 90 f.; R. Rasajski,
Rad vo/v. muzo x, 1961, 23 f.; on the large oppidum of Gomolavasee Radvo/v. muzo xiv, 1965,
1°9-25°;D. Dimitrijevic, Og. Zb. xii, 1969, 89; E. B. Bonis, Die splitkeltiseheSiedlungGellert-
hegy-Tabdnin Budapest(Budapest, 1969), 188-91; also Pannonia, 533-4, and most recently Zs.
Visy, MFME, 1970, 5 fr.
94 CIL vi, 32542 b, V. 9.12. c-d, V. 6-7.
95 Pliny, Nat. Hist. iv, 80.
97 Suetonius,Aug. 63; Plutarch, Ant. 63; Dio li, 22, 8; Frontinus,strat. I, 40, 4, etc.
98 Appian, bell. eiv. v, 75.
99 Ptolemy iii, 7, I; cf. 8, I.
100 Appian, II!Jrie. 22.
101 Seenote 85.
102 Velleius Pat. ii, 59, 4·
103 Seenote 97.
104 Horace,sat. ii, 6, 53.
105 Appian, Il!Jrie. 22 fr.; Dio xlix, 34-8.
106 Dio xlix, 36, 1.
107 Appian, Il!Jrie. 23.
108 Ibid., 15.
109 Ibid., 22; Strabovii, 5, 2, etc.
110 A. Mocsy, Historia xv, 1966, 5II fr.
III CIL i 2, p. 50; cf. Livy, epit. 134, 135; Florus ii, 26.
112 Dioli,23,2.
II3 Horace,carm. iii, 8, 18.
II4 Mocsy, op. cit. (note IIO).
I I 5 Dio li, 23-7.
II6 Ibid., liv, 20, 3.
117 Deduced by A. V. Premerstein,]OAIB i, 1898, 158 f., from Velleius Pat. ii, 39, 3;
Eusebius,ehron., 166h (Helm). Cf. Pannonia, 540, and J. J. Wilkes, Dalmatia (London. 1969),
61.
II8 Dioliv, 31, 3.
365
Notes to pp. 24-jJ
119 Cf. Pannonia, 535-7; A. M6csy, Acta Ant. xiv, 1966, 108 ff., and now F. Papazoglu,
Sred1!iobalkanskaplemenau predrimskogdoba (Sarajevo,1969)'
120 Ptolemyill, 10,4-5; CIL v, 1838 = ILS 1349.
121 Athenaeusx, p. 443 A-C.
123 Livy v, 34.
124 Polybius iv, 46, viii, 22; PompeiusTrogus,prol. 25; cf. G. Mihailov, Athenaeum,N.S.
xxxix, 1961, 39 f.
125 E. B. Vag6, Alba Regiai, 1960,49ff.; E. F. Petres,Fol. Arch. xvii, 1965,96ff.
126 J. Todorovic, Kelti ujugoistocno/Evropi (Belgrade, 1968).
127 Strabovii, 5,7; Expositio tot. mundi 51.
128 Athenaeusvi, p. 272.
129 xiii, 1,21.
130 Strabovii, 5, 2-4, 10; Appian, Illyric. 14, 22.
IF Cf. Justin xxiv, 4: ibi domitis Pannoniis.
132 The basicwork is K. Pink, 'Die Miinzpragungder Ostkeltenund ihrer Nachbarn',Diss.
Pann. ii, 15 (Budapest,1939).
133 O. Gohl, Num. Kijzl. xxi-xxii, 1922-3, 3 ff.; M. Parovic-Peiiikan,Starinar iV/II, 1961,
41 f.
134 Pink, op. cit. (note 132).
135 V. Ondrouch, Ndlezy keltskych, antickych a ryzantskychminci na Slovensku(Bratislava,
1964), 194, and E. Kolnikova, Slov. Arch. xii, 1964, 402 ff.
136 For finds of dies togetherwith tools for striking coins, seeO. Gohl, Num. Kazl. vi, 1907,
47 ff. More recently B. Benadik, Germaniaxlill, 1965, 87.
137 v, I, 8; Pannonia, 681.
138 J. Sasel,'Contributo aHa conoscenzadel commerciocon gli schiavi norici ed illirici aHa
fine del periodorepubblicano',in Atti del iii CongressoInternaz. di Epigrajia greca e latina (Rome,
1959), 143 ff.
139 C. 3776, 3777, 3780, 10721, etc.; cf. J. Sasel,PWRESuppl. xi, 561-2.
140 A. M6csy, Num. Kazl. lx-lxi, 1961-2,25ff.
141 Pannonia, 691.
142 Ibid., 692; B. Mitrea, Ephemeris Dacoromana x, 1945, 85 ff.; Parovic-Peilikan, op. cit.
(note 133),41; MS 257; LuJ i, 1961,93.
CHAPTER 2
I Illyric. 15.
2 Strabov, I, 8.
3 J. Sasel,in Corolla memoriaeE. Swobodadedicata (Cologne-Graz,1966), 198 ff.
4 Livy xliii, 5.
5 C. 3776, 3777; cf. J. Sailel, PWRESuppl. xi, 561-2.
6 Mon. AnO'r. 30; Dio liv, 34,4.
7 Cf. Pannonia, 539,46-60.
8 Dio 1, 24, 4.
9 Tacitus, Ann. 1, 80; vi, 39; Dio lviii, 25, 4.
367
Notes to pp. 33-9
10 Summaryof different views given by J. Szilagyi, Laur. Aqu. i, 304 ff. Even Mommsen
consideredthat Augustusadvancedonly as far as the Drave.
II Dio liv, 24, 3.
12 Ibid., liv, 28, I; Velleius Pat. ii, 96; Florus ii, 24.
13 Josephus,Ant. xvi, 4, I; Suetonius,Tib. 7; Gardthausen,Augustusund seineZeit ii, 668.
14 Velleius Pat. ii, 96.
15 Ibid.; Suetonius,Tib. 9; Frontinus, Strat. ii, I, 15; Dio liv, 31; 33, 5; Rufius Festus,
brev. 7, cf. Velleius Pat. ii, 39; Livy, epitome141; Eutropius vii, 9.
16 Dio liv, 34; 36, 2; lv, 2, 4; Eusebius,chron. 167 f., 168b (Helm); Cassiodoruschron.,
Chron. Min. ii, 135.
17 Mon. Anryr. 30; Dio liv, 34,4.
18 Tacitus, Ann. iv, 44; Dio lv, I la, 2, cf. Tacitus, Germ. 41; Suetonius,Nero 4.
19 Tacitus, Germ. 42; cf. Velleius Pat. ii, 108.
20 Tacitus, Ann. ii, 62.
21 Tacitus, Germ. 41.
22 Mon. Anryr. 30; Dio liv, 36, 2.
23 ILS 8965; cf. Pannonia, 543 f.
24 Florus ii, 24; Velleius Pat. ii, 96.
25 Floms ii, 28-9.
26 Strabovii, 3,10 and 13; Consolo ad Liviam 385; Florus ii, 28-9; Dio lv, 30,4, etc.
27 Dio lv, 29, 3; Strabovii, 3, 10; Florus ii, 29.
28 E. Stein, 'Die Legatenvon Moesien', Diss. Pann. i, 10 (Budapest,1940), 16.
29 Tacitus, Ann. iv, 5: ripamque Danuvii legionumduae in Pannonia,duae in Moesia attincbant. Of
thesefour legions only two, XV Apollinaris andV Macedonica,were stationedon the Danube,
at Carnuntumand Oescusrespectively.
3I Floms ii, 28-9.
F They are not mentioned on the fragmentaryinscription ILS 8965. Jordanes,Get. 74,
possiblyrefers to their location at that time, i.e. west of the river Alutus (Olt) and southof the
Carpathians.
33 Lucan, Phars. iii, 94; Seneca,quaest.nat. i, praef., 9.
34 Pliny, Nat. Hist. iv, 80.
35 For the earliestIazyganfinds, see A. M6csy, Acta Arch. iv, 1954, 124 ff.
36 Velleius Pat. ii, IIO.
38 The main sourcesfor the rebellion are Velleius Pat. ii, 110 ff., and Dio lv, 28 ff.; lvi,
1,11-17. Cf. also Suetonius,Tib. 16; Orosius vi, 21, 27, etc.
39 MS 4 8 n. 13.
40 Pannonia, 526-7.
41 ClL i 2 , p. 248 = x, 6638, col. ii, v. 3.
42 CIL xiv, 3613 = ILS 918.
43 This follows from Velleius Pat. ii, II2, 2 and II6, 2.
44 For first time ClL x, 6225 = ILS 985 and ClL xvi, 14.
45 On the cohortesBreucorum, see J. E. Bogaers,Berichten van de Rijksdienstvoor Oudheid-
kundig Bodemonderzoek xix, 1969, 27 ff.
368
Notesto pp. }9-44
46 Bevijlk. 12I •
47 CIL xvi, 20; Acta Arch. ix, 1958,407fr.
48 SuetoniusTib. 9: Pannonico(bello) Breucoset Sarmatassubegit.
49 Dio lv, 30, mentionsRomanallies in the vicinity of Alma mons = FruskaGora.
50 Dio lv, 30.
5I On the basis of the so-calledDrususinscription found at Aquincum it was unanimously
agreedto datethe immigration of the Iazygesto between17 and 20. SeePannonia,549. The new
reading, however, shows this to be wrong. See Chapter4 (p. 80).
52 Velleius Pat. ii, 123, I; cf. Suetonius,Tib. 21.
53 Tacitus, Ann. i, 16-29; Dio lvii, 4, 4; Velleius Pat. ii, 125, 4; cf. J. Sasel,Historia xix,
1970, 122-4·
54 J. Saseland 1. Weiler, Carn.-]b. 1963-4, 40.
55 It is not known where from. Formerly the deduction in Emona was linked with the
evacuationof the legionary fortress there. Cf. B. Saria, Laur. Aqu. i, 245 fr. Cogentreasons
for doubting that there was a fortress at Emona have now been put forward by J. Sasel,
PWRESuppl. xi, 562 fr., who inclines to the view that the rebellion in A.D. 14 took place at
Siscia: Historia xix, 1970, 123.
56 Tacitus, Ann. ii, 44, 46.
57 Tacitus, Ann. ii, 62 fr.; iii, I I; Velleius Pat. ii, 129, 3; Suetonius,Tib. 37; CIL xiv, 244.
The situation within the kingdom of Vannius, Pliny Nat. Hist. iv, 80-1, is controversial,cf.
Pannonia, 549; R. Hanslik, PWREviii A, 346 f.; B. Saria,ibid., 338 fr.; 1. Bona, Acta Arch. xv,
1963,303;J. Fitz, Alba Regiaii-iii, 1963,207f.; iv-v, 1965,77fr.; J. Dobias, DCU 364 fr.
58 Tacitus, Ann. xii, 29-30.
60 SeeChapter3.
61 CIL xiv, 3608 = ILS 986, cf. D. M. Pippidi, Contributi la istorie vechea Romaniei(Bucharest,
1967), 287 fr.
62 Tacitus, Hist. ii, 85 f.; 96; iii, 1-5,24; Josephus,Bell. iv, II, 2.
63 Seeprevious note and Ann. Ep. 1966, 68.
64 Tacitus, Hist. i, 79; iii, 24.
65 Ibid., iii, 5.
66 Ibid., iii, 46; iv, 4; cf. ii, 83.
67 Josephus,Bell. vii, 89-91; Tacitus, Hist. iv, 54.
68 Josephus,Bell. vii, 91-5.
69 C. 10878 = AI] 371; C. 10879 = AI] 381; C. 4060 = 10869 = AI] 260; AI] 262.
70 Tacitus, Ann. i, 16-31.
71 J. Sasel,PWRESuppl. xi, 563.
72 Dio lv, 29, 4.
73 B. Gerov, Acta Ant. xv, 1967, 85 fr.
74 Florus ii, 30 •
75 Tacitus, Ann. iv, 5.
76 Appian, Ilfyric. 30.
77 E. Stein, 'Die Legatenvon Moesien', Diss. Pann. i, 10 (Budapest,1940), 19.
78 MS 68 f.
79 ZA xii, 1962, 365; C. 8250; cf. MS 95, n. 47.
369
Notes to pp. 44-JI
80 Tacitus, Ann. i, 20. The Via Geminafrom Aquileia over the Alps was probably built
soon after 35 B.C.: J. Sasel,PWRE Supp!. xi, 573.
8I The distribution of pre-imperialcoins in particularprovesthat this was a communication
route. Cf. note 142 in previouschapter(p. 367), and A. Alf6ldi jun., Magyar Muzeum1946, 52 ff.
82 G. Alf6ldy, Acta Arch. xvi, 1964, 247 ff.
83 Tacitus, Ann. xii, 30.
84 Proof that both legions took part in the building work comes from C. 138I 3 b =
ILJug. 57 and C. 1698 = ILJug. 60.
85 A detailed description is given in E. Swoboda's scholarly work, 'Forschungenam
obermoesischenLimes', Schriften der BalkankommissionAntiquarische Abteilung x (Vienna,
1939)'
86 ILJug. 55; 58: iter Scorjularum; cf. the road-stagingpostad Scrofulason the TabulaPeuting-
eriana,eastof Viminacium.
87 Strabovii, 3, 13·
88 ILJug. 57; 60.
89 Ibid., pp. 31-9'
90 See MS 52 for a provisional view.
91 In this descriptionI dependmainly on E. Swoboda,0p. cit. (note 85).
92 For examplethe diploma CIL xvi, 4 from the year 60 lists sevencohorts in Pannonia;
inscriptions on the other hand attestroughly the samenumberof alae in the province.
93 Ri:imische Ceschichtev, 21, 178. Mommsen,however, was not so precisein his wording as
were his successors.
94 L. Barkoczi and E. B. Bonis, Acta Arch. iv, 1954, 129 ff.
95 Only brief preliminary reports:E. B. Vago, Arch. Ert. xci, 1964, 255; Alba Regiaxi, 1971,
I 12 ff. (Intercisa) and E. B. Thomas,Arch. Eri. xci, 1964, 257 (Matrica).
96 Bevi:ilk. 243, No. 158/1-8.

97 C. 4269.
98 E.g. C. 10514; 143498 ; 15163.
99 C. 3322.
100 C. 3271.
101 C. 3256 could perhapsrefer to Malata (Banostor). The strangeway of counting miles
a Malata Cusumon C. 3700-3 (Nerva) could be regardedas evidencethat at an early stageMalata
had becomea junction.
102 D. Gabler, Arrabona ix, 1967, 21.
103 C. 4269.
R~.
104 G. Juhasz,'Die Sigillaten von Brigetio', Diss. Pann. ii, 3 (Budapest,no date), 176 ff.
r05 K. Sz. Poczy, Bp. xvi, 1955, 41, Abb. 1.
106 See note 47. The establishmentof an imperial frontier on the Danubeunder Claudius
mentionedin Aurelius Victor, Caes.4, 2, will probablyhaveto be takento refer to the establish-
ment of the provincesof Noricum and Thrace.
107 E.g. A. Betz, RUG xviii, 1937, 54, 74 ff.
108 C. 4227; 4228;4244 (cf. D. Gabler, Arrabona xi, 1969,44,No. 39); Ann. Ep. 1965, 161.
109 J. Fitz, Corsium (Szekesfehervar,1970), Fig. 7; cf. on ala ScubulorumB. Gerov, Acia
Ant. xv, 1967, 95 ff.
110 ZA xii, 1962, 365; seealso C. 8250.
370
Notes to pp. JI-60
I II Lisry Filologicke vi/lxxxi, 1958, Eunomia 37.
112 C. 8261 = ILS 2733; C. 14589, cf. MS II8, n. 89.
I I 3 Tacitus, Ann. xii, 29: ipsaqueeprovincia lecta auxilia.
II 4 Bevoik. 121.
II5 CIL xvi, 2. 4; C. 4372 ; 4373; 4376 ; 4377.
II 6 Seenote 47 and CIL xvi, 31.
II7 Tacitus, Hist. iii, 12.
I I 8 Dio lv, 29, where, however,it is only statedthat there was to be recruitmentamongthe
Dalmatians.
CHAPTER 3
Nat. Hist. iii, 148.
2 CIL vi, 2385, I, 12.
3 On the settlementareasof the Pannoniantribes, seeBevo·lk. 15-80. On the Arabiates,Acta
Arch. xxi, 1969, 348.
5 Dio liv, 31 , 5.
6 Typical south Pannonian-Dalmatian namesare Bato, Breucus,Dases,Dasmenus,Liccaius,
Lucius. Cf. also Germus,Lascus,Madena,Racio, etc. KatiCic, SymposiumIg64, 3I ff.
7 B. Gerov, Acta Ant. xv, 1967, 91 ff. and in Antilnoe ObfCestvo (Moscow, 1967), 23 ff.;
Primus [IuN lib.] Asalusd[upl. ala] Cap(itoniana). Cf. Ann. Ep. 1912, 187: Iulius Saturio Iuli lib.
domo Haeduusmissic. ala Capitoniana.
9 Ibid., ii, 109, 5·
10 For theseimitations of denarii, see O. Gohl, Num. KoZI. i, 1902,17ff.; Bp. Reg.viii, 1904,
182; also A. AlfOldi, Karpatenbecken(seenote 27), 29 ff.; A. Mocsy, Num. Ko"ZI. lx-lxi, 1961-2,
25 ff.; Pannonia, 692.
I I Seenote 54.
12 C. 3377.
13 Further literature: 1. Bona, Acta Arch. xv, 1963, 304 ff.; S. Foltiny, Barb-Festschr.79 ff.;
K. Horedt, Acta Musei Napocensisv,1968,419ff.; Pannonia, 712,33-62;G. Reinfuss,Carn.-Jb.
1960-1,65,Nos 6-7 and 91; A. NeumannPWREixA, 79; B. Saria, PWREviiiA, 344, etc.
14 Ptolemy ii, 14, 2.
15 Itineraria Ant. 263, 7.
16 CIL vi, 32542.
17 Germaniai, 1917, 132 ff.; RossiusVitulus was praepositusgentis Onsorum.
19 C. 4149, 4224, 5421.
20 Germ. 42.
21 C. 3598 = 10552.
22 Listed in BeviHk.; cf. also Acta Arch. xxi, 1969, 342 f.
24 Germ. 28.
25 M. Szabo,Ant. Tan. x, 1963, 220 ff.
371
Notes to pp. 60-70
26 A. Mocsy, 'Die Lingua Pannonica',SymposiumIg66, 195 if.
27 Primarily Alf6ldi, 'Zur Geschichtedes Karpatenbeckensim 1. Jh. v. Chr.', Archivum
Europae Centro-Orientalis viii, 1942, and also OstmitteleuropalscheBibliothek, 1942.
28 CIL xvi, 4: IantumarusAndedunis f Varcianushad a Celtic name: contrastC. 4372; cf. p.
2.280: Bato Buli f Colapianushad an Illyrian-south Pannonianone.
29 West Celtic analogieswith Pannoniannameshave beeninvestigatedby M. Szabo,Arch.
Ert. xci, 1964, 165 if.
30 L. Nagy, Laur. Aqu. ii, 232 if.; cf. J. M. de Navascuesin Akte des iv. Internat. Kongr. fiir
gr. u. lat. Epigr. Wien Ig62 (Vienna, 1964), 281 if.
31 P. Petru, Arb. Vestn. xvii, 1966, 361 if.; and seech. I, n. 76 (p. 365).
32 Caesar,bell. Gall. i, 5, 4.
33 Justin xxxii, 3, 12.
34 The most recent study is that of J. Garbsch,Die norisch-pannonischeFrauentracht (Munich,
1965). Cf. J. Fitz, Gnomonxxxvii, 1965, 619 if.
35 MS 139 if.
36 Ibid., 83 if.
37 See,e.g., Spomeniklxxi, II6, 117, 252, 254, 273, 275, 276, 278, 280; lxxv, 212; lxxvii, 47;
xcviii, 272. A pair of broocheson the no longer extant part of a tombstoneis mentionedin
Spomeniklxxi, 520; ]OAIB vi, 1903, 32, No. 40. On the other stelaethe dressis fastenedon the
right shoulderby a button or by a round, button-like brooch.
38 C. 14507.
39 Cf. F. Papazoglu,Sred1!Jobalkanska plemena(Sarajevo,1969), 173 if.
40 CIL vi, 32°5.
41 C. 8242 = Spomeniklxxi, 254.
42 G. Alf6ldy, Acta Ant. xii, 1964, 109 if.; but cf. R. KatiCic, Godif1!Jakiii (Centarza Balkano-
loska Ispitivanja i, Sarajevo,1965), 63 if.
43 Strabovii, 3, 2, 10; 5,2; cf. Acta Ant. xiv, 1966,97,n. 100.
45 ILS 986.
46 CIL xvi, 13.
47 Pliny, Nat. Hist. iii, 149; Ptolemy iii, 9, 2.
48 Ibid., iii, 9, 3: their chef-lieu Ratiaria. ILS 1349: neighboursof the Triballi.
49 MS 25-9'
50 B. Gerov, loc. cit. (note 7).
51 Ptolemyiii, 9, 4·
52 CIL v, 1838 = ILS 1349.
53 C. 14387 = ILS 9199.
54 CIL ix, 5363 = ILS 2737-
55 Evidencethat the southPannoniancivitates had the statusof civitas stipendiariais possibly
provided by Aurelius Victor, epit. Caes. I, 7: Pannoniosstipendiariosadiecit.
56 C. 32.24.
57 Cf. S. Dusanic,Arch. Iugosl. viii, 1967,67.
58 Ann. Ep. 1937, 138.
59 Claudius: ILS 212, col. ii. Aelius Arist., Rom. 108. Antoninus Pius: CIL v, 532.
60 Citizens: C. 3546, 15134 (cf. ZA xv, 1965, 89, n. 17, and xvii, 1967, 198 if.); Burgenl.
372
Notes to pp. 7I-7
Heimatb. xiii, 1951, 3, NO.3 = xiv, 1952, 100; Historia vi, 1957,49°.For peregrini: Ann. Bp.
1937, 138; C. 1035 8.
61 Primarily, of course,the big cemeteryat Emona, seeLj. Plesnicar-Gecand S. Petruin a
forthcoming book on the cemetery.Poetovio: E. B. Bonis, 'Die kaiserzeitlicheKeramik von
Pannonieni', Diss. Pann. ii, 20 (Budapest,1942), 253 ff. Carnuntum: RLiO xvi, 1926; xviii,
1937, etc. Seealso Swoboda, Carnuntum245.
62 E.g. Bonis, loc. cit. (note 61),262-3;A. Barb in A. Radnoti, 'Die romischenBronzegeHisse
von Pannonien',Diss. Pann. ii, 6 (Budapest,1938), 177 ff. For tumuli, seenote 137 in Chapter
5 (p. 381 ).
63 Velleius Pat. ii, I I 0, 5.
65 Most recently J. Wielowiejski, Kontakty Noricum i Pannonii Z ludamiprflnoctrymi (Wroclaw-
Krakow, 1970), 32 ff.
66 Owing to the difficulty and in some casesthe impossibility of dating more closely the
epigraphic material relating to freedmen in the first century, the reader is referred to the
synopsisin Acta Ant. iv, 1956, 222 ff.
67 M. Abramic, ]OAIB xvii, 1914, 138.
68 RLiO xvi, 1926, 45 ff.
69 AI] 575.
70 Material listed in Mocsy, BevOik.
71 B. Vikic-BelanClc, Starinar iv, 13-14, 1965, 89 ff.; D. Gabler, Arch. Brt. xci, 1964,94ff.;
Arrabonavi, 1964, 5 ff.; ix, 1967,21ff.; Ant. Tan. xiv, 1967, 58 ff.; Acta Arch. xxiii, 1971, 83 ff.
72 E. B. Bonis, Die spiitkeltischeSiedlungGellerthef!J-Tabdnin Budapest(Budapest,1969).
73 H. Mitscha-Mahrheim,Mitteilungen des Vereins der Freunde Carnuntumsiii, 1950, 2 ff.; cf.
Swoboda,Carnuntum270.
74 E. Nowotny and O. Menghin, Wiener Prdhist. Zeitschrift xiii, 1926, 101 if.; xiv, 1927,
127 ff., 135 ff.
75 S. Nagy, Radvrjv. muzo ix, 1960, 112 ff.; and seech. I, n. 93 (p. 366).
76 K. Sz. Poczy, Bp. Reg. xvi, 1955,41 if.; K. M. Kaba, ibid., 272 ff.
77 ]OAIB xii, 1909, 226 ff.
78 Nat. Hist. iii, 147.
79 Ann. Bp. 1912, 8; 1913, 57.
80 B. Saria, Pannonia,viii, 1935, 171 f.
81 Hence I cannot share Swoboda'sdoubts concerningmy interpretationof the dec. Scarb.
Seehis Carnuntum242 f. Both in Scarbantiaand in Savariadonors of inscriptions, particularly
in the early years, saw to it that all the titles and namesby which the town was known were
transcribed.
82 It is unnecessaryto do more than cite J. SaikI's excellent article on Emona (PWRE
Suppl. xi), which containsall the important bibliographical sources.
83 Pliny, Nat. Hist. iii, 147; Velleius Pat. ii, 109, 5.
84 Bevijlk. No. 88/1 = C. 1°921;89/1-2; 90/5 = C. 4171.
85 Ibid., 38.
86 Ibid., No. 89/2.
87 A. Alfoldi jun., Arch. Brt. 1943, 71 ff.
88 C. 4156.
373
Notesto pp. 77-8;
89 J. Sasel,tA x, 1960, 201 if.
90 E.g. CIL v, 943; 1011. Cf. Beviilk. 9°/30: P. OpponiusIustus Cl. Savaria Aculeiensis.
91 CIL v, 8336.
92 C. 10936 = 4225; cf. ILS 8507.
93 C. 4009, 4250, 4 251, II259; Ann. Ep. 1938, 163.
94 C. 41 96.
95 E. Turr, Arch. Ert.lxxx, 1953, 130. Previousinvestigationsand finds are listed by T. P.
Buocz in Savaria topogrdfttija (Szombathely,1968). A new attempt to reconstructthe street-
network in the colonia is to be found in Arch. Ert. xcviii, 1971, 193 if., by E. T6th.
96 M6csy, Arch. Ert. xcii, 1965, 27 if.
97 C. 4224; Beviilk. 97/1.
98 C. 4212, 10926.
99 C. 4199, 4200; Bevolk. 38, n. 151.
100 Ibid., 180/1.
101 A. Mocsy, Arch. Ert. lxxxi, 1954, 167 if.
CHAPTER 4
I Josephus,bell. vii, 4, 3.
2 C. III94-7.
3 E. Toth and G. V6kony, Acta Arch. xxii, 1970, 133 if.
4 Cf. Swoboda,Carnuntum, 36.
5 C.459I.
6 Starinar iV/18, 1967,21if.
7 K. Wachtel, Historia xv, 1966, 247.
8 C. 8261 = ILS 2733.
9 Seenote 6.
10 Spomenikxcviii, 441.
II Suetonius,Domitianus6; Eutropius vii, 23, 4; Jordanes,Get. 76.
12 Most recently R. Syme, Arh. Vestn. xix, 1968, 103.
13 Suetonius,Domitianus6, 5; Jordanes,Get. 77.
14 Dio lxvii, 10.
15 The expansionof the auxiliary forces in Pannoniabetween80 and 85 can easily be traced
from the diplomas: CIL xvi, 26, 30, 3I.
16 Tacitus, Ann. xii, 30; Hist. iii, 5; 21.
17 Dio lxvii, 7.
18 ILS 9200.
19 Dio lxvii, 5.
20 Ibid., lxvii, 5, 2.
21 Suetonius,Domitianus6; Eutropiusvii, 23, 4; Tacitus, Agr. 41.
22 Martial vii, 8; viii, 8; ix, 3I. His stay at Carnuntumis attestedby C. 4497.
23 J. Dobias, Corolla memoriae E. Swoboda dedicata (Cologne-Graz,1966), 117, n. 10; cf.
Historica iv, 1962, 27.
24 CILv, 7425 = ILS 2720. Pliny,paneg.8, 2; 16, 1.
25 Szilagyi, Inscr. teg. 69.
374
Notes to pp. 86-97
26 C. 10224; VHAD ix, 1906-7, 103, Nr. 220; seealso C. 13360.
27 C. 3468; Bp. Reg. viii, 1904, 162, No. I; C. 143572;cf. G. Alf6ldy, Acta Arch. xi, 1959,
132 ff.
28 C. 14349, 2, 4, 9.
29 C. 10513,10514,143498,15163, T. Nagy, Bp. Mueml. ii, 36.
30 A. M6csy, Acta Arch. iv, 1954, II5 ff.
31 C. 1435922 ; cf. A. Neumann,PWREixA, 67 f.
32 C. 151964= Jahrbuchdes Vereinsfur Geschichteder Stadt Wien xvii-xviii, 1961-2, 8, No. I.
33 Szilagyi, Inscr. teg. 83 f.; L. Bark6czi, Brigetio 20.
34 Rufus Festus,brev. 8.
35 L. Bark6czi and E. B. B6nis, Acta Arch. iv, 1954, 129 ff.
36 Cf. Chapter2, notes 96-101 (p. 370), and, e.g., C. 10246 = 3262, elL xvi, 54, etc.
37 C. 164 2 •
38 A. R. Neumann,Forschungenin Vindobona I948 bis I967, i, 'Lager und Lagerterritorium',
RLiO xxiii (Vienna, 1967). Bark6czi, Brigetio.
39 elL vi, 1548; Tacitus, Germ. 42.
40 V. Ondrouch,Historica Slovacaii, 1941, 22 ff., Nos II2-24, 188. T. Kolnik, LRKN 40.
41 For the buildings: Pannonia, 643 f., with bibliography. Purposeof the buildings: Acta
Arch. xxi, 1969, 355 f. The discoveryof a hitherto unknownbuilding at Paccould not be taken
into consideration:seeT. Kolnik, Arch. RoZhl. xxiv, 1972, 59 ff.
42 Dio lxviii, 10, 3
43 Oratio xii, 16-20 (transl. J. W. Cohoon).
44 CIL xvi, 39 and 46.
45 Bark6czi, Brigetio 20.
46 C. 10517. K. M. Kaba, Bp. Reg. xvi, 1955, 275; cf. Acta Arch. xxi, 1969, 352, n. III.
47 C. 15162. Bevolk. 185/33; Bp. Reg. xxi, 1964, 247.
48 CIL xvi, 47.
49 SHA, Hadrianus 3,9.
50 Doubts aboutthis, expressedby S. Dusanic,Arch. lugosl. viii, 1967,69,are in my opinion
unfounded.
51 J. Sasel,PWRESuppl. xi, 571-6.
52 This can also be deducedfrom the reliefs of Trajan's Column, which show Roxolanian
troopersin coats of mail fighting the Romans.'
53 Cf. Jordanes,Get. 75.
54 Dio lxviii 10,3; cf. Acta Arch. iv, 1954, I25.
55 SHA, Hadrianus 3, 9; Dio lxviii, 10, 3; Eusebius,chron. 194b (Helm); Eutropiusviii, 3, I.
56 A. M6csy, Acta Arch. iv, 1954, I24 ff.
57 Ptolemyiii, 7, I; 8, I.
58 A. M6csy, loc. cit. (note 56); M. Parducz,MFME 1956,15 ff.; Zs. Visy, MFME 1970,3 ff.
59 D. Protase,Acta Mus. Napocensisiv, 1967, 47 ff.
60 H. Nesselhauf,elL xvi, p. 224 ad. n. 164.
61 MS 50 ff.
62 Starinar iV/I 5-16, 1965, 173 ff.
63 CILxvi, III (A.D. 160).
64 MS 50 ff.
375
Notes to pp. 97-IOri
65 c. 8262, 14575, 14577,14579; ]DAIB xii, 183, No. 52.
66 Patsch,Beitriige V(2), 90 if.; D. Tudor, Podurile romane de la Dunarea de jos (Bucharest,
1971), 53 if.
67 Dio lxviii, 13,6.
68 MS 18-25.
69 Ptolemy, Ceogr.,passim;Itineraria Ant. 242-8; Dio lv, 23; CIL vi, 3492 = ILS 2288, etc.
70 Arch. Ert.lxxviii, 1951, 135; Bp. Reg. xvii, 1956, 165; Acta Arch. xi, 1959, 256; Bp. Reg.
xx, 1963, 27 f.
71 SHA, Hadrianus 6, 6.
72 Ibid., 6, 6; cf. R. Syme,JRSIii, 1962, 87.
73 SHA, Hadrianus 6,7; 7, 3·
74 CIL v, 32, 33 = ILS 852, 853·
75 On the war, see SHA, Hadrianus 5, 2; Eusebius,cbron. 198d (Helm); Orosius vii, 13, 4,
Epigrap~y,
and C. and H. Daicoviciu, Acta of the Jth International Congressof Creek and Latin
CambridgeIg67 (Oxford, 1971), 347.
76 Dio lxix, 15,2.
77 Caes. 13, 3·
78 P. Lakatos, Ant. Tan. xii, 1965, 91 if., and MFME 1964/5, 65 if.
79 Most recently N. Gostar,SCIVii, 1951, 2, 169 f.
80 Seenote 59.
81 SHA, Hadrianus 9, I; Eutropius viii, 6, 2.
82 L. Barkoczi, Intercisa ii, 504.
83 E. Jonas,Bp. Reg. xii, 1937, 287 if.
84 SHA, Hadrianus 12, 7.
85 C. 6818 = ILS 1017, cf. J. Dobias in Omagiu lui C. Daicoviciu (Bucharest,1960), 147 if.;
Historica iv, 1962, 28 f.; Dobias connectshim with the wars under Hadrian. This cannotbe
right as legio XIII Geminahad no legionarylegateat that time. To connecthim with the wars
under Domitian would be a possibility.
86 Listed in Pannonia, 554; cf. J. Fitz, Acta Ant. xi, 1963, 255, and L. Balla, Acta Classica
Debreceni, 1965, 44 if.
87 SHA, Aelius 3, 2; cf. Hadrianus 23, 12.
88 CIL xi, 5212 = ILS 1058.
89 RIC iii, 620.
90 SHA, Marcus 12, 13; cf. R. GobI, RheinischesMuseumciv, 196I, 70 if.
91 R. Noll, Archaeologia Austriaca xiv, 1954, 43 if.; cf. E. Swoboda, Carn.-Jb. 1956, 5 if.;
Pannonia, 555; J. Fitz, Acta Ant. xi, 1963, 262, 266; Nagy, Bp. Miieml. ii, 41 f.
92 R. GobI, Zwei riimische Miinzhorte, IlImitz und Apetlon (Eisenstadt,1967).
93 Tac: J. Fitz, Acta Arch. xxiv, 1972, 3 if. Doboj: Arh. Pregl. viii, 1966,122; ix, 1967,93.
The fort of Doboj in the valley of the Bosnawas perhapsa frontier stationon the Pannonian-
Dalmatianprovincial boundary.
94 There are several detailed studies of the history of Pannonianarmy troop-movement
in the secondcentury. SeeespeciallyL. Barkoczi and A. Radnoti, Acta Arch. i, 195I; T. Nagy,
Acta Arch. vii, 1956 and J. Fitz, Acta Ant. vii, 1959.
95 Nagy, Bp. Miieml. ii, 519 if.
96 Itineraria Ant. 244, 4.
376
Notesto pp. I06-I}
97 Swoboda,Carnuntum15. T. Nagy's recentexcavationsof Aquincumhaveshownthat even
the legionaryfortress there had to be moved back (seeFig. 23).
98 Pannonia, 634-8.
99 D. Gabler, Arrabona viii, 1966, 67 if.; Arch. Brt. xciv, 1967, 221; xcv, 1968, 130 f.; xcviii,
1971, 269.
100 See Chapter2, note 95 (p. 370).
101 Arch. Brt.lxxxii, 1955,67if. Cf. also Nagy, Bp. Miieml. ii, 93, n. 51, on first-centuryfinds
under the paving stones.
102 Gabler, op. cit. (note 99); cf. Acta Arch. xxi, 1969, 354, n. 138.
103 M. v. Groller, RLiO iv, 1903, 23 if.; S. Soproni,Arch. Brt.lxxxvii, 1960, 234; Bark6czi,
Brigetio 6.
104 'Forschungenam obermosischenLimes', Schriften der Balkankommission,Antiquarische
Abteilungx (Vienna, 1939).
105 C. 3700; Burgenl. Heimatb. i-ii, 1932-3,77,No. 82; C. 4641-53, etc.
106 Prooim. 7.
107 Rom. 80-4.
108 lLJug. 63; Swoboda,'Forschungenam obermosischenLimes', 78 if.; H. U. Instinsky,
JOAlB xxxv, 1943, 33 if. For the new Trajanic inscription, see P. Petrovic, Arch. lugosl. ix,
1968, 83 if. On the canal, Procopius,aedif. iv, 6, 8-16; cf. MS II6, note 71.
109 E.g. Swoboda,Carnuntum, 271.
110 Bp. Tort. i, 76I, Abb. 36.
III B. Swoboda,Slov. Arch. x, 1962,422if.
I I 2 RLiO i, 1900, 17 if.; cf. Swoboda,CarnuntutJt 253.
I I 3 Bp. Tort. i, 748, Abb. 3I; cf. Pannonia, 641.
II4 L. Nagy, AZ Eskii-#ri romai erb'd, Pest varos ose (Budapest,1946).
115 T. Nagy, Acta Ant. vi, I958, 429 f.
II6 A. M6csy, Fol. Arch. x, 1958, 96 if.
II7 Aquincum (Budapest,District xiii): 89 by 67 m. Lussonium:100 by 55 m. Lugio: 85 by
59 m.
II8 G. Juhasz,'Die Sigillaten von Brigetio', Diss. Pann. ii, 3 (Budapest,no date), 196 if.
II 9 J. Szilagyi, Bp. Rifg. xviii, I 958, 53 if.; Bp. Miieml. ii, 352 if.; R. Egger, 'Das Praetorium
als Amtssitz und Quartier romischer Spitzenfunktionare',Sitzungsberichteder Osterreicbiscben
Akademieeel, I966, 28-36. Should Egger's assumptionbe correct, the governor'spalaceat
Carnuntumwas also on the bank of the Danube.On the dateof the palaceat Aquincum, see1.
Wellner, Arch. Brt. xcvii, 1970, 116 if., who showstwo main periods,the first of Trajanic date,
later destroyed,and the seconddating to the middle of the secondcenturyor later.
CHAPTER 5
1 Nat. Hist. iii, 148.

2 ClL xvi, 18.
3 S. Dusanic,Arch. lugosl. viii, 1967,78,n. 59, and R. Syme,Arh. Vestn.xix, 1968, 103, date
the founding of Sirmium to the reign of Domitian, without good groundsin my opinion.
4 ClL xvi, 14.
5 Tacitus, Hist. iii, 12.
377
Notes to pp. I I}-22
6 C. 3971.
7 elL xvi, 15.
8 E.g. C. 3896, 3898, 3917, 3922, 392,8, 392,9, 3932., 3935,4009,10802" 108°5,108°9,10812"
1082,4,14°42,and 1435422;AI] 489, 5°1.
9 Only brief preliminary reportsin Arh. Pregl. Also Dj. Mano-Zisi, ZNMB iv, 1964, 93 if.,
and M. Parovic-Pesikan,Starinar iV/15-16, 1966,31if.; iv/19, 1969,75if.; iv/2,O, 197°,2,65if.
10 Dio xlix, 37, 3.
II A. F. Marsigli, DanubiusPannonico-Mysicusii (Hagaeet Amstelodami,172,6),47f.; Tab. 2.0,
Fig. x. C. Veith, 'Die FeldziigedesC. lui. Octavianusin Illyrien', Schriftender Balkankommission,
AntiquarischeAbteilungvii (Vienna, 1904), 51 if. AI], p. 2,37 f.
12, C. 10865 = 4001 = AI] 588; C. 3685 = 102,49; elL xvi, 18, etc.
13 Ann. tip. 1911, 2,37.
14 ZA iv, 1954, 196, No. 2,; MS 68 f.
15 ]GAIB xiii, 2.16, No. 2,9; Spomeniklxxi, 650; cf. MS 67, nn. 40-I.
17 MS 68 if. For further details, ibid., 62,-75.
18 N. Vulic, Nekolikopita'!Ja iZ antickeistorije naIe zemije(Belgrade,1961), 87 if.; C. Truhelka,
GSND v (2,), 192,9, 78 if.
19 ]GAIB xiii, 2,18, No. 31.
2,0 MS 30 n. 9.
2,1 Listed in MS 68.
2,2, AI] 374, 375; C. 40 56, 4057.
2,3 C. 40 57 = AI] 373.
2.4 AI] 375.
2.5 B. Saria, PWRE xxi, II67 if.; 1. Mikl-Curk, Arh. Vestn. xv-xvi, 1965, 2,59 if.; for the
bridge, see AI] 361.
26 C. 8087, 142,17 = 145°0; ]GAIB xxxi, 106, lNo. II (cf. MS 107, n. 66); IBAI xiv,
194°-2" 2,72,·
2,7 See the tombstonesfrom the oldest cemetery-roadat Poetovio: M. Abramic, ]GAIB
xvii, 1914, 138; cf. A. M6csy, Acta Arch. iii, 1953, 181.
2.8 Gromatici veteres(ed. Lachmann)i, pp. 204 f.
2,9 C. 32,79. StephanusByz. s. v. Mursa.
30 NoctesAtt. xvi, 13, 5.
31 D. Pinterovic, Os). Zb. v, 1956, 55 if.; xi, 1967, 2,3 if.; Lu]i, 1961, 35 if.
32, For details, seeBeviilk.
33 For the Canii, seeC. 10901 = 3689; Bevolk. 82,/1; C. 42,5°;Ann. tip. 1912,8 = 1913, 57;
C. 3599; Fol. Arch. ix, 1957, 83 if. For the Caesernii,seeJ. Sasel,ZA x, 1960, 201 if.
34 Bevolk. 74/1-2,; cf. 90/30'
35 Ibid., 73/2,; C. 102,89 = 3308 = Bevoik. 2,16/5.
36 Bevo"lk. 186/II.
37 C. 11047·
38 Bevolk. 164/8,2,4,2.5.
39 SeeChapter3, note 71 (p. 373).
40 Bark6czi, Brigetio 37; for new evidence,see Acta Ant. xiii, 1965, Plate xxiv.
41 T. SzentleIeky,Bp. Reg. xix, 1959, 197.
378
Notes to pp. I24-}4
42 A. Mocsy, Fol. Arch. ix, 1957,83ff.; cf. also xvi, 1964,43ff.
43 D. Laczko and Gy. Rh<::, Baldcza(edited by K. Hornig, Veszprem,1912).
44 Arch. Brt. lxxx, 1953, Taf. xvii, 28.
45 C. 6480 = 10954 = Arch. Brt. xlii, 1928, 207 if.
46 CIL v, 8973.
47 C. 4122; cf. E. Ritterling, PWRExii, 1252; C. 4123, 4148, 143556•
48 C. 415 3, 4499; cf. Beviilk. 164/12, 14·
49 Full length: T. Nagy, Bp. Reg. xiii, 1943,464f. Eagle: C. 143498 = SchoberNo. 57.
50 L. Nagy, Germaniaxvi, 1932, 290 f.
51 C. 10548; L. Nagy, Germaniaxv, 1931, 263; xvi, 1932, 288 ff.
52 Beviilk. I 86/II.
53 Ibid., 186/30 : domo Nemes= the singular of Nemetes(J. Szilagyi).
54 Ann. Bp. 1933, IIO.
55 C. 10430 .
56 C. 3505.
57 Barkoczi, Brigetio No. 35·
58 Carnuntum:RLiG vii, 1906, 95 ff.; viii, 1907, 7 ff.; ix, 1908,43ff. Aquincum: Bp. Miieml.
ii, TaE. vi. Vindobona: A. R. Neumann,Forschungenin Vindobona i (see Chapter 4, note 38
(p. 375). Brigetio: Barkoczi, Brigetio 20.
59 Carnuntum:RUG vii, 1906, 83 ff. A similar building at Aquincum (J. Szilagyi, Bp. Reg.
xiv, 1945, 133; Bp. Miieml. ii, 518) was situated,as recentexcavationsby T. Nagy haveshown,
within the walls of the legionaryfortress.
60 C. 4500, 4501, 11301, 13379,1435515, 143594 ; Ann. Bp. 1929,217;1939,261,etc.
61 For the time being, seeB. Rutkowski, Acta Rei Cretariae RomanaeFautorum, v-vi, 47 if.;
cf. MS 259 f.
62 MS 200, Abb. 45.
63 MS 148 if.
64 C. 14505.
66 CIL v, 1047 = ILS 7526.
67 MS 191 f.
68 M. Vasic, ]ahrb. DAI xx, 1905; Archiiol. Anzeiger102 if.; MS 145 f.
69 MS 126.
70 ]GAIB vi, 23, No. 30.
71 R. Mowat, Revuede Numismatiqueiii (12),1894,372ff.; cf. K. Regling, PWRExv, 1322 if.
72 Aedif· iv, 4, 3·
73 Itinerarium Ant. 134, I; Itinerarium Burdig. 565, 1 and Tab. Peut.
74 O. Davies, RomanMines in Europe (Oxford, 1935), 209 if.
75 M. Velickovic, ZNMB I, 1958,95 ff.
76 C. 6313 = 8333; 14536.
77 C. 14606; Spomeniklxxi, 251.
78 Spomeniklxxi, 217; cf. MS 35.
79 The hitherto unpublishedresults of Emil Cedkov's excavationswill provide detailed
information on the metalla Dardanica. This energeticrepresentativeof the youngergeneration
of Serbianarchaeologistswas unfortunatelykilled in a roadaccidentin 1969.In 1965 he showed
379
Notes to pp. I3J-42
me some of the results of his highly successfulexcavationsat Ulpianum and Socanica.See
now E. CerSkov,Municipium DD (Pristina-Belgrade,1970)'
80 Historia vi, 1957, 490 if.
81 Burgenl. Heimatb. xiii, 1951, 3, No. 103; xiv, 1952, 100.
82 SeeBevijlk. II8/1, 130/1, 131/1-2, 132/1, 133/2, 136/4,14, 15.
83 elL ix, 5363 = ILS 2737; on the dating, see K. Wachtel, Historia xv, 1966, 247.
84 C. 3679, 1°783;AI] 234.
85 C. 3919, 3921, 3925, 10804. Itinerarium Ant. 259, 13.
86 C. 3679.
87 C. 3896, 3917, 3922, 3928, 3929, 3932, 10791, 10802, 108°5, 108°9,108II, 14042 and
1435422.
88 C. 4°°9,10824, 10866,II463; AI] 489.
89 C. 3925, 10801.
90 BeviJlk. 24/2, 25/1, 32/1, 33/1, 34/2-3, 35/1-2, 36/1, 37/1, 38/1.
91 C. 3546.
92 E.g. Ulpii: C. 3375, 3407, 3410, 10334, 10377, 143416, 15151. Aelii: C. 10355, 10408,
10993, II043; BeviJlk. II7/1; Ann. tip. 1953, 14.
93 Athenaeusvi, p. 272D (from Agatharchides).
94 MS 86f.
95 Spomeniklxxi, 182,276;lxxvii, 3I; ]OAIB vi, 40, No. 46. For further details, seeMS 83.
96 A certain Rujinus Dast" pr(inceps?) is mentioned on an inscription from the Metohija,
Spomeniklxxi, 278. This is possibly the only epigraphicevidenceof a dignitary of the civitates
peregrinaeof Upper Moesia.
97 elL xii, II22.
98 C. 3676; Dio lxix, 9, 6.
99 Intercisa i, No. 294.
100 Aquincum: L. Nagy, Bp. TiM. i, 372, 467; Laur. Aqu. ii, 191; T. Nagy, Bp. Reg. xxi,
1964,9if., and especiallyActa Arch. xxiii, 1971, 59 fr. Carnuntum:Swoboda,earnuntum154.
101 At present,seeA. Mocsy, Stud. Mil. 211 if. and Acta Ant. xx, 1974.
102 C. 4554; Ann. tip. 1938, 167.
103 C. 143592 •
104 C. 8127, 8128, 12659, 138°5,142172.Spomeniklxxi, 308, 311.
105 C. 1655.
106 C. 3347, 10334, 10355, 10377, 143416. Intercisa i, No. 294. Ann. tip. 1953, 14; 1965,
12.
107 C. 10430, 10525. Intercisa i, No. 25. Ann. tip. 1933, IIO; 1939, 8-9' BeviJlk. 185/10;
186/II, 30,43'
108 C. 10418; cf. A. Alf6ldi, Arch. tirt.lii, 1939, 108.
109 Fol. Arch. xxi, 1970, 59 if. J. Fitz holds, though without good grounds,that Gorsium
was a municipium: seeActa Arch. xxiii, 1971,47if.; xxiv, 1972, 3 if.
110 C. 10408; cf. A. AlfOldi, op. cit. (note 108).
III Alba Regia xi, 1971, 127, No. 460.
112 elL xvi, II2, 123, 179, and diploma of A.D. 157 to be publishedby Zs. Visy.
113 Ann. tip. 1953, 14.
114 Seenote 108.
380
Notesto pp. I42-54
1I5 C. 10336; Ann. Bp. 1953, II.
II6 C. 10305.
1I7 G. AlfOldy, Epigraphica xxvi, 1965,95 ff.
1I8 elL vi, 3297.
1I9 E.g. the attack on Plautius Silvanus' troops by the Pannoniansin A.D. 7, Gallienus'
victory over Ingenuusin 258 or 259, Constantine'svictory over Licinius in 314 and the battle
betweenConstantiusII and Magnentiusin 351. For recentexcavationsat Cibalae,seeB. Vikic-
BelanCic, Vjesnik arheolofkogmuzejau Zagrebuiv, 1970, 159 ff.
120 S. Dusanic, Arch. Iugosl. viii, 1967, 67 ff.
121 Seenote 117.
122 C. 3267.
123 S. Dusanic,ZA xv, 1965, 85 ff.
124 C. 10993, 1I043.
125 C. 10900, 151881.
126 BevOlk. 53 f.
127 C. 4267, 4490.
128 Information kindly suppliedby M. Gorenc.
129 Cf. A. Mocsy, Arch. Brt. xci, 1964, 16 f.
130 MS 75 f.; E. l:erSkov, Rimijani na Kosovtl i Metohiji (Belgrade,1970)'
IF Spomeniklxxi, 204; lxxv, 161; xcviii, 222.
132 MS 86, Abb. 27.
133 E.g. J. Csalog, Arch. Brt. 1943,41 ff.; Gy. Torok, Arch. Brt.lxxvii, 1950, 4ff. Cf. also
I. Hunyady, 'Die Kelten im Karpatenbecken',Diss. Pann. ii, 18 (Budapest,1944), 152 f.; M.
Parducz,Acta Arch. ii, 1952, 143 ff.; iv, 1954,25 ff. etc.
134 Listed by K. Sagi, Arch. Brt. 1944-5, 214ff.; lxxviii, 1951, 75.
135 There is no exhaustivestudy: see Pannonia, 723 f.; also M. Seper, Arheolofki radovi i
raspraveii, 1962, 335 ff.; A. Radnoti, Bcryerische Vorgeschichtsbla'tterxxviii, 1963, 67 ff.
136 Fol. Arch. xiv, 1962, 35 ff.
137 K. Sagi, Arch. Brt. 1943, 113 ff.; Pannonia, 718 f., also H. Kerchler, Die rijmerzeitlichen
Brandbestatttlngentinter Hugel (Vienna, 1967); M. Amand, Latomtls xxiv, 1965,614ff; A. Barb,
Gnomonxl, 1968, 501; E. B. Bonis, Fol. Arch. xiv, 1962,23ff.; S. Pahic,Starinarvii-viii, 1958,
310; Arh. Vestn. xi-xii, 1960-1, II6 ff.; A. Neumann,Forschtlngenin Vindobonaii, 30 ff.
138 Tombstonesof the native population are arrangedaccordingto place of discoveryin
A. Mocsy, Bevlilk. Seemap, ibid., which is the basisof the map (Fig. 26) in this volume.
139 E.g. C. 4224; cf. also A. Barb, Btlrgenl. Heimatb. i, 1932,78,No. 87; xiii, 1951, 216 ff.;
xxii, 196o, 166 ff.
140 C. 10895.
141 T. P. Buocz, Vasi Szemle1962, I, 107.
142 SHA, Marctls 21, 7.
143 MS 166-78.
144 C. 8239; Spomeniklxxi, 253, 280.
145 Spomeniklxxv, 141 = xcviii, 179.
146 Seein particular C. 14507; cf. MS 172 ff.
147 E.g. C. 4371 and 4367; but seealso C. 4368, which commemoratesa Batavianin the ala
I Ityraeorum.
381
Notesto pp. IJJ-JI
148 Forinstance,RLiO xii, 321; xviii,43, No.8; 61, No. 22; Ann. Ep. 1929, 212; C. 4473,
143582Ia and 4491. Cf. R. Syme, Arh. Vestn. xix, 1968, 108.
149 E.g. Ann. Ep. 1929,2°5,206,208,2°9,220; 1934, 174; 1937, 267.
150 E.g. C. 3528, 3530,10497-500.
151 Pannoniusalso of courseoccurs as indication of the place of birth; but there is some
evidencethat a legal implication wasinvolved. Pannoniansoldiersdischargedfrom the exercitus
Pannonicuswereindicatedby the adjectiveof the tribe or civitas, whereasthosedischargedfrom
non-Pannonianarmies were indicated by the province. The only exceptionwas a Boian dis-
chargedfrom the exercitusRaetiaeand mentionedin CIL xvi, 55.
152 C. 10897: a Tib. Claudius . .. marif
153 C. 3679'
154 Ann. Ep. 19°9,235= 1938, 13·
155 C. 14507.
156 Ulcisia Castra, Albertfalva, Campona, Matrica, Vetus Salina, Intercisa, Annamatia,
Lussonium, AltaRipa, Alisca(?).
157 The origo castris has no connectionwith birth in the canabae; see Acta Ant. xiii, 1965,
4 2 5-31'
158 See,e.g., the list in C. 14507 or the namesin legio II Adiutrix, Bevo'lk. 185/7, 10, 13, 15,
16, 17, 18, 19,20,21,25,etc.
159 Seefollowing note.
16o Diplomas found in Pannoniaand their find-spots can be divided as follows: from the
Claudianand Flavian periods seven,including one or two from fort-sites (CIL xvi, 26, 30?)
and five elsewhere(CIL xvi, 2, 17, 18, 20 and 31). BetweenA.D. 96 and 167 thirty-one diplo-
mas,including nineteenfrom fort-sites (CIL xvi, 47, 49, 61, 69, 71, 73, 77, 89, 97, 99, 109, II3,
II6, II9, 123, 132, 164, 175; Acta Arch. ix, 413) and twelve elsewhere(CIL xvi, 42,64, 84, ...
92,96, 100, 103, 104, 112, 178, 179, 180).
161 For what follows, seeBp. Tort., Bp. Miieml. and the relevantsectionof the bibliography
(p. 425). On the wells of the aqueductsee K. Sz. P6czy, Arch. Brt. xcix, 1972, 15 ff.
162 Swoboda,Carnuntumis the standardwork.
163 Swoboda,Carnuntum83 ff.
164 A. Betz, Carn.-Jb. 1960 (1962),29ff.; H. G. Kolbe, ibid., 1963-4(1965),48ff.
165 Treatedin detail by Swoboda,Carnuntttm 154 ff.
166 The lateststudy is that of P. Petru,T. Knez and A. Ursie, Arh. Vestn. xvii, 1966, 491 ff.
167 M. Grbic, Antiquiry x, 1936, 275.
168 K. Sz. P6czy, Arch. Ert. xciv, 1967, 137 ff.
169 D. Boskovic, Starinar iii/4, 1928, 270, Abb. 1.
170 M. Vasic, Jahrb. DAI xx, 19°5;Archiiol. Anzeiger102 ff.
171 Catnuntum:L. Klima and H. Vetters,RLiO xx, 1953.Aquincum: T. Nagy, Bp. Reg.xiii,
1943, 368 ff.; Bp. Miieml. ii, 405 ff.
172 BaLica: D. Laczk6 and Gy. Rhe, Baldcza(edited by K. Hornig, Veszprem,1912). Eisen-
stadt: W. Kubitschek, RO'merfunde21 ff. Gyulafirit6t-Poganytelek: Gy. Rhe, Os-es dkori
tryomok Veszpremko'riil (Budapest,1905).
173 Parndorf: B. Saria, Barb-Festschr.252 ff.
174 Deutschkreuz:A. Barb, Geschichteder Altertumsforschungim Burgenland(Eisenstadt,1954),
19. Donnerskirchen:Kubitschek, RO'merjunde49. Regelsbrunn:RLiO iii, 1902, 14 if.
382
Notesto pp. I7I-9
175 Winden am See: B. Saria, Der riimische GutshofvonWinden am See(Eisenstadt,1951)'
176 Bahica, Parndorf, Donnerskirchen, Regelsbrunn and Smarje-Grobelce:F. Lorger,
Casopisza Zgodovinoin narodopiije xxix, 1934, 147 if.; xxxi, 1936, 77 if.
177 St Georgen:A. Barb in Radn6ti, Bronzegefiisse,Diss. Pann. ii, 6 (Budapest,1938), 194.
Fertorikos:D. Gabler, Arch. Brt. xcii, 1965, 235; xciii, 1966, 294, and probablyalso Winden
am See(seenote 175)'
178 E. B. B6nis, Die spatkeltischeSiedlung Gellerthegy-Tabdnin Budapest (Budapest, 1969),
119-36.
179 E. Kocztur, unpublishedexcavations.See meanwhileJ. Fitz, Gorsium (Szekesfehervar,
1970), PI. 43.
180 E.g. J. Pdkar, Pam. Arch. Iii, 1961, 414 if.; J. Pavlik, Arch. RoZhl. xvi, 1964, 323 if.;
M. Lamiova-Schmiedlova,Slov. Arch. xvii, 1969, 402 if.; T. Kolnik, Slov. Arch. xix, 1971,
50 7.
181 A. M6csy, Arch. Brt.lxxxii, 1955, 62 if.
182 Gy. Torok, Fol. Arch. xiii, 1961, 63 if. For the burials from this site, see E. B. B6nis,
Fol. Arch. xii, 1960, 91 if. Alternatively, thesehousescould have beenbuilt with turf (a sug-
gestionof R. Muller).
183 Seenote 181; also T. Nagy, Bp. Miieml. ii, 37, 50 f. (Albertfalva); E. B. Vag6, Arch. Brt.
xci, 1964, 255 (Interdsa),etc.
184 J. Hampel, Arch. Brt. v, 1885, 24 if. The diploma: CIL xvi, 96. The bronzevessels:A.
Radn6ti, Bronzegefiisse,Taf. xxii, 5; xxv, 1-2.
185 On Tac (= Gorsium), see most recently J. Fitz, Gorsium (Szekesfehervar,1970), and
Acta Arch. xxiv, 1972, 3 if.
186 This is attestedby tiles stampedTE PR found there and by a votive inscription set up by
a sacerdoscalled L. Virius L. f. Mercatorpro salutetemplensium.Virius was possiblya native of
Sirmium (cf. L. Virrius Iustus Sirmiensisat Aquileia, S. Panciera,La vita economicadi Aquileia,
1957,76).
187 L. Barkoczi and E. B. B6nis, Acta Arch. iv, 1954, 150 if.; E. B. Thomas, Acta Arch.
vi, 1955, II4 if.
188 E. Bonis, 'Die kaiserzeitlicheKeramik von Pannonieni', Diss. Pann. ii, 20 (Budapest,
1942); K. Sz. P6czy, Acta Arch. xi, 1959, 151 if.; E. B. Vag6, Alba Regiai, 1960, 54 if., etc.
189 Pannonia, 676 f.
190 Pannonia,679 f.; also E. F. Petres,Fol. Arch. xvii, 1965, 96 if.; B. Vikic-Be1anCic,Starinar
iV/13-14, 1965,97if.; G. Reinfuss,Carn.-Jb. 1960 (1962), 78 f.
191 Pannonia, 679, cf. D. Gabler, Arrabona vi, 1964, 5 if.
192 In an area south-eastof Sala Roman sites do not yield coins or sigillata: see R. Muller,
RegesZetiterepbfJdrdsok(Zalaegerszeg,1971), 79 if.
193 B. Kuzsinszky,Bp. R~. xi, 1932; K. Kiss, Laur. Aqu. i, 212 if.; K. Sz. P6czy,Acta Arch.
vii, 1956, 102 if.
194 A. AlfOldi, Fol. Arch. i-ii, 1939, 97 if.
195 L. Nagy, Bp. Reg. xiv, 1945, 305 if.; B. Rutkowski, Acta Rei Cretariae RomanaeFautorum
x, 1968, 18 if. A bowl mould found at Gorsium is also attributableto the so-called 'Sisda'
pottery, but does not prove that Gorsium was the centreof this industry.
196 E.g. SchoberNos 120, 159, 170, 171 and 267.
197 E.g. ibid., Nos III, 113, 125, 152, 186,209,210and 246.
383
Notes to pp. I8o-7
198 Provisional referencemay be made to MS 62-160, Abb. 20, 24, 29, 32, 35, 37, 39, 41
and 42.
199 A. Sz. Burger, Bp. Reg. xix, 1959, 9 ff.
200 Scarbantia:C. Praschniker, ]OAI xxx, 1937, I I I ff. Savaria:1. Paulovics,Arch. Erl. 1940,
34 ff.
201 Pannonia, 744; MS 243 ff.
202 C. 4156, 10908 (Isis), 10913 (Sphinx). From ScarbantiaC. 4234, cf. V. Wessetzky,Das
Alterlum x, 1964, 154 ff.
203 A. Betz in CarnuntumI33J-I9}J (Vienna, 1935),28ff.
204 C. 4418; RLiO xii, 1914, 321 f.
205 AI] pp. 133 ff. At Aquincum the Mithraeum datesto 161-3: C. 3479.
206 E.g. C. 3416,4°°9, 10395.
CHAPTER 6
I SHA, Marcus 12, 13.

2 ILS 8977; cf. CIL vi, 1497 = ILS 1094, etc. On other units, Saxer, Vexillationen 33; A.
Radn6ti, Germaniaxxxix, 1961, 109; T. Nagy, Bp. Miieml. ii, 94, n. 108, etc.
3 E. Ritterling, P WRE xii, 1544f.; tile-stamp: Bp. Reg. xvii, 1956, 105.
4 The participationof legio VII Claudiais attested,seee.g., ]OAIB viii, 19, NO.5 8.
5 Cf. H. G. Pflaum, Essaisur les procurateursequestres(Paris, 1951),71f.
6 SHA, Marcus 24, 5; 27, 10; cf. Dio lxxi, 33,42•
7 F. Lorger, ]OAI xix-xx, 1919, 107 ff. For the history of II Italica, seeG. Winkler, ]ahrbuch
des oberbsterreichischenMusealvereinscxvi, 1971, 86 ff.
8 iii, 12; cf.ii, 17; iv, 12.
9 De pallio, ii, 7.
I I SHA, Marcus 14, I; aliis etiamgentibus, quae pulsae a superioribusbarbarisfugerant, nisi reci-
perentur, bellum inferentibus.
12 For some archaeologicalevidence of new peoples,see J. Tejral, Pam. Arch. lxi, 1970,
184 ff.
13 Barbarianchieftains wished to be incorporatedinto the empire: Appian, preface,7.
14 Dio lxxi 3, Ia.
15 Cf. SHA, Verus 9, 9.
16 Seenote I 1.
17 SHA, Marcus 22, I: gentesomnes... conspiraverant.
18 Lucian, pseudomant.48; Ammianus Marcellinus xxviii, 6, I.
19 SHA, Marcus 14,2; Verus 9,7-10; Galen xiv, 649 f.; xix, 17 f.
20 SHA, Marcus 14, 6.
21 ILS 8977.
22 SHA, Verus 9,9-10; Marcus 14, 3.
23 CIL iii, pp. 213 f.
24 This would be so if ClaudiusFronto, legateof Moesia,was simultaneouslycomesof Verus.
On the situation in Dacia, see L. Balla, Acta Classica Univ. Scient. de L. Kossuth nominatae
Debreceniensisvii, 1971, 73 ff.
384
Notes to pp. 187-93
25 CIL vi, 1377 = ILS 1098.
26 A. Birley, SeptimiusSeverus(London, 1971), II3.
27 DoM, Verwaltung, 64 ff.
28 Spomeniklxxv, 2.
29 EspeciallyC. 6302 = ILS 2606; cf. MS 16, n. 50.
30 Dio lxxi, 3, 5.
31 Dio lxxi, 3, 12; Eutropius viii, 13, I; Eusebius,chron. 207a (Helm).
32 CIL vi, 8878 = ILS 1685; Philostratus, vito soph. ii, I, 26 ff. A moving documentis
dated18 March 175-thewill of a young Italian who died at Sirmium, probably of the plague
still raging there, after all his servantshad died saveone slave namedAprilis: CIL x, 7457.
33 Dio lxxi, 8-10; Eusebius,chron. 206i (Helm); cf. the forged letter of Marcus in R. Merkel-
bach, Acta Ant. xvi, 1968, 339 ff.
34 Dio lxxi, II, 1-6. Furtius: Dio lxxi, 13, 3.
35 Ibid., lxxi, 12, 1-2.
36 Ibid., lxxi, 13, 3-4.
37 Ibid., lxxi, 15.
38 Libyca iii (1955), 145 = Ann. Bp. 1956, 124.
39 Cf. the affair of TarruntenusPaternus,Dio lxxi, 12, 3.
40 Dio lxxi, 12, 3.
41 Ibid., lxxi, 12, 3, on their defeat. Cives Cotini from round Mursa and Cibalaeservedin the
praetorianguard under the Severi, see CIL vi, 32542.
42 Ibid., lxxi, 21.
43 Ibid., lxxi, 7.
44 Ibid., lxxi, 15.
45 lxxi, 13, 1-2.
46 lxxi, 14, 1.
47 lxxi, 16, 1.
48 lxxi, 16, 2.
49 lxxi, 17·
50 CIL vi, 1599 = ILS 1326; cf. SHA, Marcus 13,5.
5I Dio lxxi, 33; SHA, Commodus12, 6.
52 Dio lxxi, 18; 19,2.
53 Ibid., lxxi, 20; CIL viii, 619 = ILS 2747; C. 13439 = ILS 9120; Ann. Bp. 1956, 124.
54 lxxi, 20.
55 lxxii, 3, 2.
56 Dio lxxii, 15, 3; Eutropius viii, 15, I; Aurelius Victor, Caes. 17, epit. 17; SHA, Com-
modus II, 8; 12, 7.
57 Ann. Bp. 1956, 124.
58 Dio lxxii, 2.
59 Germaniai (1917), 132 ff.
60 For a new summaryof the material, see D. Gabler, Bqyerische Vorgeschichtsbliitterxxxiii,
1968, 100 ff.; Arch. Brt. xcv (1968), 2II ff. See also J. Tejral, Arch. RoZhl. xxii, 1970,
389 ff.
61 Dio lxxi, 20, 2; cf. 33, 4 2 •
62 SHA, Marcus 24, 5; 27, 10; Herodiani, 5, 6.
385
Notes to pp. I9J-200
63 Dio lxxii, 2, 4.
64 Ibid., lxxi, II, 2; 13,4; 16,2.
65 Pannonia, 562; J. Fitz, Acta Arch. xiv, 1962, 76; T. Nagy, Bp. Miieml. ii, 44 if.; A. Neu-
mann, PWREixA, 78, etc.
66 L. Bark6czi, Intercisa ii, 519; A. M6csy, Arch. Brt.lxxxii, 1955,65if.; Eireneiv, 1965, 143,
n. 147; T. Nagy, Bp. Miieml. ii, 101, n. 287; J. Fitz, Acta Arch. xiv, 1962, 76.
67 CIL viii, 619,14605,16553, 25740, 25894, 27512; iii, 3680,4472,10419, 10515, 13372;
J. Fitz, Alba Regiavi-vii, 1966, 208; Ann. Bp. 1956, 124.
68 C. 3324, 3444, 3542, 3545, 3675; Arch. Brt.lxxxii, 1955, 62 f.
69 The extensiveinformation about this unit is given in Intercisa i-ii; for the arrival of the
cohort, seeJ. Fitz, Acta Ant. xi, 1963, 277.
70 C. 14542,6302= 8162 = ILS 2606; cf. MS 16, n. 50.
71 C. 142176, 14537, 14541,14545; jOAIB xii, 189, No. 59; Starinar ivf5-6, 358.
72 N. Vulic, jahrb. DAI xxvii, 1912; Archa'ol. Anzeiger, 549 if.; xxviii, 1913, 339 if.; xxix,
191 4, 4 12•
73 C. 14537·
74 SHA, Marcus 21, 7.
75 MS, 110 if.; cf. 194 if.
76 C. 825 I; Spomeniklxxi, 253; cf. MS 195.
77 C. 3385; Intercisa i, Nos 297-307.
78 SHA, Commodus6, I; cf. also Dio lxxii, 8, I.
79 Herodiani, 9, I.
80 SHA, Commodus13, 5. For the disgraceof the governorCorneliusFelix Plotianus,seeJ.
Fitz, PWRESuppl. ix, s.v. Cornelius No. 284.
81 SHA, Commodus12,8; CIL v, 2155 = ILS 1574.
82 Herodianii, 9, 12; SHA, Sevcrus5, 3.
83 An inscription (Arch. Brt. lxxviii, 1951, 135) establishesthe presenceof legio illr Flavia
at Aquincum in the governorshipof Q. Caecilius Rufinus Crepereianus(Dob6, Verwaltung,
78) which can be dated c. 202-6, since legio II Adiutrix was definitely back in Pannonia
by 207; M. Mirkovic, ZFF vii (i), 1963, 1I 3 if. On the absenceof this legion under Caracalla,
seeM. Mirkovic, ZA xi, 1962, 319 if.
84 C. 3660; Intcrcisa i, No. 326; cf. G. Alf6ldy, Arch. Brt. lxxxviii, 1961, 29 f.
85 Pannonia, 587 f.
86 Dio lxxvii, 20, 3-4.
87 Herodianiv, 7, 3-5; 8, I.
88 Dio lxxvii, 20, 3-4.
89 C. 10505; 143495 ; Arch. Brt. 1944-5, 178.
90 Seenote 41.
91 See,for example,Herodianvi, 7.
92 P. Dura 54, col. ii, 3.
93 E. Ritterling, PWRExii, 1686. There may be a connectionbetweenthe initial reluctance
of X Geminaand the troubles in Noricum, cf. CIL ii, 4114 = ILS 1140.
94 Ann. Bp. 1941, 166; cf. A. Betz, in Corolla memoriacE. Swobodadedicata (Cologne-Graz,
1966), 39 if.
95 Herodianii, 13; cf. Dio lxxv, 2; cf. A. Passerini,Le coorti pretoric (Rome, 1939), 174 if.;
386
Notes to pp. 200-9
Pannonia,646; M. Pavan,AthenaeumN.S., xl, 1962,85if.; A. Dobo, 'Inscriptionesad res Pan-

nonicaspertinentes',Diss. Pann. i, I (Budapest,1940), No. 1-62, 144a-5.
96 i, 9; ii, 8-II, 13-15, etc.
97 Appian, I1!J1ric. 6.
98 Dio lxxx, 4-5.
99 Princ. hist. 13.
100 Herodianii, 9; cf. Dio xlix, 36,2.
101 E.g. Dio lxxv, 2, 6.
102 SHA, Severus10, 3.
103 CIL viii, 7062 = ILS II43. The young Caracalla was entrustedto the governor of
Pannonia,Fabius Cilo, who was an intimate friend of the Dynasty; cf. Dio lxxvii, 4, 2.
104 Herodian iii, I, 10; cf. J. Fitz, Acta Arch. xi, 1959, 237 if.; F. Papazoglu,Bulletin de
CorrespondanceHellenique lxxxv, 1961, 162 if.
105 Cf. G. Alfoldy, Acta Ant. vi, 1958, 192. The Pannonianswere the most loyal of the
supportersof Maximinus Thrax, accordingto Herodianvii, 8; viii, 6.
106 Herodianvi, 7.
107 Both hoardsin the hinterland of the ripa Sarmatica: S. Soproni, Acta Arch. xvii, 1965,
275 if.
108 Herodianvii, 2, 9; SHA, Maximinus Thrax 13, 3.
109 Its importanceis alreadyemphasizedby Herodianvii, 2, 9.
IIO SHA, Aurelianus 3, I; cf. under Carus, SHA, Carus 4, 3.
III E. Jonas, OMRTE i, 1923-6, 137 if.; A. Radnoti, Num. KoZI. xliv-xlv, 1945-6, 6 if;
K. B. Sey, Num. KoZI. lxiv-Ixv, 1965-6, 9 if.
II2 Zosimus I, 21, 2.
I I 3 For thesesenior commanders,see J. Fitz, Stud. Mil. I I 3 if.
II4 Zosimusi, 21, 2-3.

II 5 CIL ii, 2220; viii, 1430.
II6 Listed by K. B. Sey, Arch. Ert. xcviii, 1971, 199; cf. the map, ibid., 196, fig. 6.
I 17 Eutropiusix, 8, 2; SHA, trig. tyro 10, I; Panegyrici Latini viii (Baehrens),10, 2, etc.
II8 For example,Starinar iii/6, 32 if.; IBAI xxii, 360.
119 MS 257 f.
120 SHA, trig. ryr. 9; Chron. Min. i, 521; Aurelius Victor, Caes. 33, 2.
121 Zonarasxii, 24; Eutropiusix, 8, 1.
122 SHA, trig. ryr. 10; Aurelius Victor, Caes. 33, 2; epit. 32, 3.
123 R. GobI, Der ro'mischeSchatzfundvon Apetlon (Eisenstadt,1954); cf. Pannonia, 569.
124 Aurelius Victor, Caes. 33, I; epit. 33, 1.
125 C. 3228; cf. p. 2382'82.
126 J. Sasel,'Bellum Serdicense',Situla iv (Ljubljana, 1961).
127 It becamevirtually a commonplaceamonglate Romanhistoriansthat the candidatewas
found to becomeemperoragainsthis will and only after threatsto his life.
128 Most recently J. Fitz in MelangesCarcopino (Paris, 1966), 353 if. There is no explanation
for the absenceof somelegionesfrom the seriesof coin-reverses,VII P(ia) and VII F(idelis).
129 Most recently E. Pasalic,Stud. Mil. 127 if.; cf. Pannonia, 662 f., 666.
130 M. Mirkovic, RGD, 66 f.
131 Jordanes,Get. xvii; cf. 1. Bona, in OroshdzaTo"rienete, i (Orosh:iza, 1965), 114.
387
Notes to pp. 209-IJ
132 SHA, trig. {yr. 9; cf. J. Harmatta,Studiesin the History of the Sarma/ians(Budapest,1950),
60 f.
133 AI], pp. 144 ff.
134 Cf. C. Daicoviciu, La Tran.rylvanie dansl'Antiquite (Cluj, 1945), 184 f.; cf. C. Daicoviciu,
Dacica (Cluj, 1970), 378, and D. Tudor, Historia xiv, 1965, 369 ff.; Historica i (Bucharest,
1970),67ff.
135 SHA, Gallienus 13; Claudius 6, II; Zosimus i, 43, 45, etc.; cf. A. AlfOldi, CAH xii,
149 f., 72 I ff.
136 Claudius: SHA, Claudius 13,2. Aurelianus: SHA, Aurelianus3,1-2; Eutropiusix, 13, 1.
Heraclianushas a name whose Latin variant appearsfrequently in Upper Moesia; see A.
M6csy, Godifrljak viii (Centar za BalkanoloskaIspitivanja vi, 1970), 160 f.; for the cult of
Hercules in Moesia Superior, see MS 245; and now a votive stone to Omphale: Starinar
iV/19, 1969, 225.
137 SHA, Aurelianus 3, 1.
138 Aurelius Victor, Caes. 39, 26; J. Straub,in Vom Kaiseridealder Spatantike(Berlin, 1938),
209, explains the remarkablecombinationof satis with a superlativeby saying that optimus,
the normal adjectivefor good emperorsafter Trajan, had lost its superlativecharacter.A more
probableexplanation,however, for which I thank ProfessorI. Borzsak, is that satis appears
frequently in later Latin with the meaning of 'as possible'. Cf. also Aurelius Victor, satis
commodans,40, 9; satis clarus, 32, 2 (the comparableinstanceoccursin epit. 32, I with splendidis-
simus).Satisoptimuscomescloseto the expressionnecessariusmagisquambonusin SHA, Aurelianus
37, 1.
139 SHA, Aurelianus 18, 2; 30, 5; Zosimus i, 48-9; Dexippus, frg. 7; cf. A. Alfoldi, Serta
Kazarovianai (Sofia, 1950),21ff.
140 For examplethe bath-houseof the Aquincum legionaryfortress was rebuilt (c. 3525 =
1°492),and the third Mithraeum at Poetovio was built (note 133 above), etc.
141 Ammianus Marcellinus xxxi, 5, 17; cf. SHA, Aurelianus22, 1-2.
142 Panegyrici Latini xi (Baehrens),17; Jordanes,Get. II 6-20.
143 Eutropiusix, 15, 1.
CHAPTER 7
I The reforms of Commodushave been studied by J. Fitz in severalarticles; see especially
Klio xxxix, 1961, 199 ff.
2 P. Petrovic, Starinar iV/18, 1967, 57 ff.
3 ClL vi, 2386.
4 Spomeniklxxi, 248; lxxv, 168.
5 MS 35, 43, 90 ff.
6 J. Deininger, Die Provinziallandtageder riimischen Kaiserzeit (Munich, 1965), 119; cf. MS
92 •
7 MS 42 ff.
8 C. 1674-6, 8244, 8249, 12672 = 14561; Spomeniklxxvii, 37.
9 Spomeniklxxi, 594.
10 MS 142 f.; 159.
II Kanitz, RS69 f.; B. Saria, BRGKxvi, 1925-6, 93; D. Piletic, Arh. Pregl. iv, 1962, 176 f.
388
Notes to pp. 2I5-2}
12 C. 7591; ClL vi, 2388.
13 jOAlB xii, 1909, 23, No. 30.
14 Starinar iV/I, 1950, 143 ff.; iV/2, 1951, 115.
15 C. 8141.
16 MS 144f.
17 MS 134 ff.
18 C. 6302 = 8162 = lLS 2606.
19 MS 134 ff.; 226 ff.
20 Spomeniklxxi, 584.
21 C. 14217';jOAlB vi, 60, No. 99; xiii, 223, No. 39; XV, 236, No. 39; ZNMB iv, 1964, 127;
cf. jOAlB xii, 188, NO.5 8, etc.
22 C. 14541I; cf. MS 140, n. 56.
23 Both were coloniae Septimiae:seeC. 14347, 143593 ; RLiO xvi, 1926, 117. On the date 194,
see Pannonia, 599: contra A. Betz, Carn.-jb. 1960 (1962), 31, who believes that the colony at
Carnuntumwas probablyfoundedin 198. But whereasAquincum certainlyexistedas a colony
in 198, its lustrum years(i.e. the yearsin which the quinquennalesheld office) fell in the years4 and
9 of any decade(e.g. A.D. 214, 259) as C. 10439, 10440 show us. Thus it cannot have been
foundedin 198, and the date must be pushedback to 194-thefirst availabledatein the reign
of SeptimiusSeverus.
24 MS 31.
25 C. 14509; cf. MS 145, n. 2.
26 A. Mocsy, in Stud. Mil. 2II fr.
27 Cf. A. Mocsy, Germaniaxliv, 1966, 312 fr.
28 Birth in the canabaecould be includedon inscriptionsas a juridically meaninglessindication
of the birth place, e.g. ClL vi, 32783.
29 C. 3973, 3976, 4193 and 10836.
30 C. 4152; cf. Pannonia, 602.
3 I See,e.g.,lLS 6090 = Fontesiuris Romanianteiustinianii Z, No. 92: iisdemmaximepol/icentibus,
quod ... decurionumsufftciensfutura sit copia. ...
32 The only caseis that of the civitas Eraviscorumon an altar of the time of Philip (seeA.
Alfoldi, Arch. Ert.Iii, 1939, 266). But this civitasperegrinahad alreadybeenin practicedissolved
under Hadrian and existed only in nameafter 124.
33 Cf. A. Mocsy, Acta Ant. x, 1962, 367 fr.
34 C. II007·
35 Barkoczi, Brigetio, No. 226.
36 C.4557·
37 L. Barkoczi, Acta Arch. iii, 1953, 201 f.
38 ltinerarium Burdig. 561, 10.
39 Egger, RAFCi, 57 ff.
40 ClL vi, 2494a; 3241.
41 C. 3974.
42 C. 10243.
43 lntercisa i, NO.1 24.
44 Cf. Beviflk. 74.
45 MS 33 f.; M. Mirkovic, RGD 37 fr.
388
Notes to pp. 22}-}O
46 Complete list of finds: D. Garasanin,GodiI'?fak Muzeja Grada Beograda i, 1954, 60 if.;
cf. Mirkovic, RGD 48 f.
47 C. 8297; cf. MS 34 f.; E. Cedkov, Municipium DD (Pristina-Belgrade,1970).
48 C. 14610.
50 Procopius,aedif. iv, 4, 3.
5I Ibid.
52 MS 38 f., 110 if.
53 M. Dimic, Starinar viii, 1891, 21 if.; J. Mirkovic, Starinar v, 1888,72if.; MS III.
54 MS 158 f.; Mirkovic, RGD 68 f.
55 Itinerarium Burdig. 564, 5.
56 Codex Theod. i, 32, 5 = Codexlust. xi, 7, 4.
57 E. Cedkov, op. cit. (note 47), 64, No. I I.
58 MS 141 f.
59 MS II7 if.; Mirkovic, RGD 89 if.
60 C. 14038 (Cibalae); 10197 (Bassiana).In CIL vi, 32542b, 9. 17 Cibalae belongs to the
pseudo-tribusAurelia, indicating a promotion, but one which I considerimprobable under
Marcus.
61 Mirkovic, RGD 65 f.
62 C. 4335.
63 Cf. A. Mocsy, Acta Ant. x, 1962, 367 if.
64 C. 8089, 12646, 8263+14580;cf. MS 107.
65 C. 10495, 10496; Ann. Ep. 1968,423.
66 Intercisa i-ii (Budapest, 1954-7). See also J. Fitz's recent book Les Syriens a Intercisa
(Brussels,1972).
67 C. 143592 •
68 C. 10597.
69 MS 191 f.; cf. V. Velkov, in Etudes Historiques du xii e Congres International des Sciences
Historiquesa Vienne, I96; (Sofia, 1965), 25 if.
70 For example C. 10920, II076; Barkoczi, Brigetio, No. 95; ZA x, 1960, 193; ]DAIB xii,
1909, 158, No. 22, etc.
71 C. 2006; Z. Kadar and L. Balla, Ant. Tan. vi, 1959, II2 if.; L. Balla, Acta Arch. xv, 1963,
225 if.; I. Toth, Arch. Ert. xcviii, 1971, 80 if.
72 D. Pinterovic, Od. Zb. ix-x, 1965, 72 f.

73 Intercisa i, No. 329·
74 S. Scheiber,Corpus Inscriptionum Hungariae Iudaicarum (Budapest,1960).
75 L. Nagy, Arch. Ert. Iii, 1939, 116 if.
76 C. 4281, 10481, 10533, 10570; Barkoczi, Brigetio, Nos 93, 102, 209.
77 E.g. C. 3490, 11076, 10982,15167; cf. PWRE Suppl. xi, s.v. Toleses;Bp. Reg. xiv, 1945,
541.
78 Intercisa i, Nos 38, 138; C. 4220.
79 E.g. C. 3490; Barkoczi, Brigetio Nos 93, II 3; Kadar and Balla, op. cit. (note 71).
80 For exampleC. 11076.
81 For exampleC. 1°515, 14349.
82 G. Alf6ldy, in EpigraphischeStudieniv (Cologne-Graz,1967), 26 if.
390
Notes to pp. 2}0-4I
83 A. Alf6ldi, Zu den SchicksalenSiebenbiirgensim Altertum (Budapest,1944), 39 if.
84 Cf. the cataloguein BeviJlk. and L. Bark6czi, Acta Arch. xvi, 1964; also the map in MS
200, Abb. 45.
85 MS 67.
86 For Pannonia,seeD. Gabler, Arch. Ert. xciii, 1966, 20 if.
87 Cf. MS 133, 154, 174.
88 ClL xiii, 1766; SHA, Severus10, 7; cf. G. Alf6ldy, Acta Ant. viii, 1960, 145 if.
89 For exampleC. 3349, 3628, 3659, 4146, 4282, 10347, 151884,etc.
90 For the evidenceof inscriptions,seeDob6, Verwaltung 159 if. For the benejiciarii in Upper
Moesia, seeMS 22, Abb. 7 and 24.
91 C. 4146, 4268,4375, 10956, etc.
92 C. 6322, 12666-8; Spomeniklxxi, 594, 595.
93 C. 8258 = 1689; 1684 with Spomeniklxxi, 243; C. 14555, 14556.
94 MS 99 f., 143·
95 M. Vasic, Jahrb. DAl xx, 1905; Archiiol. Anzeiger102 if.
96 SeeGabler, op. cit. (note 86).
97 Cf. the papers by H. Kenner, A. Kiss and D. Mano-Zisi in La mosai"quegreco-romaine
(Paris, 1965), 89 if., 290 if., 297 if.
98 L. Nagy, RiJmischeMitteilungen xl, 1925, 51 if.; Bp. Reg. xiii, 1943, 79 if.
99 I. Wellner, Acta Arch. xxi, 1969, 235 if.
100 See,for example,L. Nagy, RijmischeMitteilungen xli, 1926,79if.; K. Sz. P6czy, Bp. Reg.
xvi, 1955,49if.; xviii, 1958,103if.; T. Nagy, Bp. Reg. xviii, 1958, 149 if.; H. Brandenstein,
Carn.-Jb. 1961-2 (1963), 5 if., etc.
101 MS 152.
102 B. Filow, lBAD i, 1910, 8 if.; cf. MS 104.
103 See, for example,Bark6czi, Brigetio 40 if.; G. Erdelyi, lntercisa i, 199 f.; Pannonia, 720;
A. Cermanovic-Kuzmanovic,Arch. lugosl. vi, 1965, 89 if.
104 SeeChapter 5, note 177 (p. 383).
105 D. Stricevic, Starinar iV/7-8, 1958,411if.
106 A. M6csy, Arch. Ert. lxxxii, 1955, 62 if.; G. Alf6ldy, Arch. Erl. xc, 1963, 302; S.
Soproni, Arch. Ert. xciii, 1966, 296; T. Nagy, Bp. Miieml. ii, 37, 50 f.; E. B. Vag6, Arch. Ert.
xci, 1964,255'
107 See, e.g., the map in K. Sagi and K. Bakay, Epi#s is Epiteszettudomdny ii, 1971, 406.
108 T. Nagy, Arch. Ert. lxxxii, 1955, 97; Bp. Reg. xvii, 1956,9;L. Nagy, Mumienbegriibnisse
3 f.; K. Sz. P6czy, Bp. Reg. xxii, 1971,98if. A large house,probablya villa, has recentlybeen
excavatedby A. Sz. Burger at Koml6 north of Sopianae,dating from Severantimes, Janus
PannoniusMuzeumEVkiJ'nyve, 1967,61if. (but seeChapter9, note 18 (p. 399».
109 S. Gaddy, Arch. Ert. xlix, 1936, 88 if.
IlO C. 10521, 10536.
III C. 11076, 12659, 142174;Bark6czi, Brigetio, No. 114.
Il2 See,e.g., C. 4311,10962,11076; cf. Bark6czi, Brigetio, Nos 91-127.Seeon the otherhand
the sarcophagiin Upper Moesiawhich belongedto the highestsocial order: MS 104, 130, 152.
113 For exampleJOAlB xiii, 1910, 203, No. 13; cf. MS 168.
114 C. 8169, 8238 = JOAlB vi, 38, No. 44; 8240 = Spomeniklxxi, 326; Spomenikxcviii,
450; JOAlB vi, 28, Nos 33-4.
391
jVotes to pp. 24I-6
II 5 FragmentVat. 220.
II6 MS II7 ff., and notes 93-4 above.
117 Aurelius Victor, Caes. 39, 26; cf. Chapter6, note 138 (p. 388).
II8 See,e.g., the excavationby Bark6czi, Acta Arch. xvi, 1964, 273 f., 278; C. 3464, 3478,
3515, 3539, 35 61, 3571, 3572, 4 28 5, 4 286, 4 299, 4327, 1°389, II082 and 143493 ; Bark6czi,
Brigetio, Nos 141, 208, 216, etc.
119 A. Kiss, Acta Arch. xi, 1959, 159 ff.
120 Aurelius Victor, Caes. 29, I; epit. 29, I; Eutropiusix, 4.
121 SHA, Probus 21,2; Aurelius Victor, Caes. 37,4; cf. epit. 37, 1.
122 At presentI can only refer to the ample unpublishedmaterial, e.g., in the Museum of
SremskaMitrovica.
123 V. Velkov, Eirene v, 1966, 174.
124 Philostratus,vito soph. ii, 56o.
125 T. SzentIeleky,in Neue Beitrag,ezur Geschichteder Alten Welt ii (Berlin, 1965), 381 ff.; Das
Isis-Heiligtum von Szombathe!J(Szombathely,1965); cf. A. M6csy, Acta Arch. xxi, 1969, 370,
n. 331.
126 C. 3925,4495;CIL vii, 341; xvi, 179-80.
127 Dio lxxviii, 13, 4. He was descendedfrom a Celtic family (Triccus > Triccianus) whose
citizenshipwent back to Hadrian or to Antoninus Pius (Aelius).
128 Ti. ClaudiusMarinus Pacatianus,Q. MessiusTraianusDecius, M. Aemilius Aemilianus,
P. C( ...) Regalianus,L. Domitius Aurelianus, C. Valerius Diocletianus, etc. On the other
hand Claudius Gothicus,Probus,Maximianus and others were Marci Aurelii.
129 D. Gabler, Arch. Ert. xci, 1964, 94 ff.; Bcryerische VorgeschichtsbMtterxxxi, 1966, 123 ff.
130 E. T. Szonyi, Acta Arch. xxv, 1973, 87 ff.
131 L. Bark6czi, Fol. Arch. xviii, 1966, 67 ff.
132 I. Sellye, 'Les bronzesemailles delaPannonie',Diss. Pann.ii, 8 (Budapest,1939).
133 N. Lang, Bp. Reg. xii, 1937, 5 ff.
134 Ibid., x, 1923, 3 ff.; Z. Kadar, Fol. Arch. xiv, 1962,41 ff.
135 Radnoti, Bronzegefiisse,82 if., 105 if., 137 if.; Intercisa ii, 190 f., 212 f.
136 E.g. B. Kuzsinszky, Bp. Reg. x, 1923,60f.; E. Swoboda,Carn.-Jb. 1955 (1956), 15 if.
There are now two further examplesof the glass cameo with the inscription MNHMONEYE
(Swoboda,op. cit.), namelyD. Tudor, Romula(Bucharest,1968), Abb. 16, and an unpublished
examplefound at Brigetio, now in the HungarianNational Museum.
137 For exampleE. B. Thomasin Studienzur Geschichteund Philosophiedes Altertums, edited
by J. Harmatta(Budapest,1968), 337 ff.; L. Nagy, Mumienbegrabnisse,18 f., etc.
138 B. Filow, IBAD i, 1910, 8 ff.; Ferri, ARD 366 ff.; Archeologija(Sofia), vi (1964), i, 24 ff.;
MS 104; Spomenikxcviii, 317.
139 Seenote 130.
140 For exampleI. Kovrig, 'Die Haupttypender kaiserzcitlichenFibeln in Pannonien',Diss.
Pann., ii, 4 (Budapest,1937), Taf. xxvi, 4; xxvii,!.
141 Pannonia, 688; J. Fromols, Jahrbuch des Rijmisch-GermanischenZentralmuseumsv, 1958,
259 ff.; J. M. C. Toynbee,Art in Britain under the Romans(Oxford, 1964), 167 f.; cf. also Acta
Arch. xxi, 1969, 358 f.
142 Pliny, Nat. Hist. xvi, 7, 43-4; Lucan, Pharsalia vi, 220; Martial, xiii, 69; Nemesianus,
cynegetica126; CIL vi, 10005 = ILS 5285; cf. also Acta Arch. xxi, 1969, 359.
392
Notes to pp. 240-54
143 Expositio tot. mundi, 51.
144 E.g. Arch. Ert.lxxxii, 1955,60,and K. Sz. Poczy, Intercisa ii, 95.
145 Except of coursetemporarydeliveries, such as the corn which the Marcomannihad to
produceunder the peace-treatyof 18o (Dio lxxii, 2, 2). This corn probably went to the army
in Pannonia.
146 Seenote 89.
147 CIL vi, 32542.
148 Compiledby A. Sz. Burger, 'Aldozati jelenetPannoniakoemlekein',RegesZetiFiizetek,ii,
5 (Budapest,1959); cf. Pannonia, 725. The sceneappearssporadicallyalso in Upper Moesia,
e.g. Starinar iV/I, 1950, 205, Fig. 53.
149 C. 14507; cf. MS 172 ff.
150 CIL vi, 32°5;cf. MS 249 f.
15 I SHA, Severus10, 7.
152 K. Wigand, ]OAIB xviii, 1915, 189 ff.; AI], pp. 154 ff.; Pannonia, 744.
153 Pannonia,741-3·
154 Cf. G. E. F. Chilver, Cisalpine Gaul (Oxford, 1941), 193.
155 1. Szanto,Arch. Ert.lxxviii, 1951,43.
156 Ferri, ARD 165, Fig. 169; L. Balla, Z. Kadar, A. Mocsy and others, Die riimischen Stein-
denkmalervon Savaria (Budapest,1971), Nos 81-4.
157 A. Dobrovits, Fol. Arch. iii-iv, 1942, 304.
158 For exampleC. 4410.
159 A. D. Nock, CAHxii, 415, n.!.
160 Magla, Messor(if not referring to harvest)and Vidasus.
161 AI], 230 ff.; cf. M. Abramic, FestschriftR. Egger (Klagenfurt, 1952), 317 ff.
162 E.g. Bp. Reg.iii, 1891, 69, 71; v, 1897, I26;ix, 1906, 47ff.; xi, 1932, 381; xii, 1932,139;
Kubitschek, Riimerfunde,104, etc.; cf. H. Kenner, ]OAI xliii, 1956-8, 91.
163 C. 3498.
164 C. 3491.
165 C. 4425-32, 13469-73.
166 E. M. Staerman,Vjestnik drevnd istorii, 1946, 3, 264.
167 C. 13368.
168 Seethe map in Zs. Banki, Alba Regia vi-vii, 1967, 165 ff.
169 Strabo vii, 5, 10; Pliny, Nat. Hist. iii, 148; Appian, Ilfyric. 14; Aurelius Victor, Caes.
40, 9; Gromatici Veteres(ed. Lachmann),p. 2°5.
170 E.g. C. 1°414;Intercisa i, No. 369; ]OAI xlii, 1955, 109 f.
171 C. 3617.
172 C. 3474.
173 Pannonia, 740, 745·
174 MS 243 ff.
175 E.g. C. 1677, 1678, 8171, 14563, 14565; Spomeniklxxi, 251; xcviii, 222; cf. MS 244 f.
176 Spomeniklxxi, 210 f., 325, xcviii, 223.
177 ]OAIB vi, 41, No. 48; xii, 172, NO.3 5; Spomeniklxxi, 191; xcviii, 171-2.
178 C. 63°3, 8148; Spomeniklxxi, 3, III, 322; lxxvii, 36; xcviii, 10, 173. Godif,!/ak Muzda
Grada Beogradaii, 36;]OAIB xv, 235, No. 38; 236, No. 40. A god with the nameAtta or Atto
may be connectedwith Paternus:seeP. Petrovic, Vraf!fski Glasnik v, 1969, 368.
393
Notes to pp. 254-62
179 C. 8184.
180 ]OAIB vi, 40, No. 46; Spomeniklxxi, 182,278,323;lxxv, 167.
lSI D. Tudor, Corpus monumentorumequitum Danuvinorum(Lei den, 1969)'
IS2 See,e.g., note 141.
183 SeeChapter5, note 205 (p. 384); also Spomeniklxxvii, 28-30 and 64; lxxv, 155.
184 M. Vermaseren,Corpus Inscriptionum et MonumentorumReligionis Mithraicae, ii (Hagae,
1960), 170 if.; cf. also T. Nagy, Arch. Ert.lxxxix, 1962,281if.
185 Z. Kadar, Die kleinasiatischenund syrischenKulte in Ungarn (Lei den, 1962),2f., shows that
the oriental cults were not widely distributed among the garrisonsof Syrian units.
186 N. Ling, Arch. Ert. 1943,64ff.; cf. Z. Kadar, op. cit. (note 185), 39 f.
187 C. 3343; cf. R. Egger,in Omagiului C. Daicoviciu (Bucharest,1960), 167 ff.; for a different
view, seePannonia, 595.
188 J. Dell, Archiiologisch-EpigraphischeMittheilungen aus Osterreich-Ungarnxvi, 1893, 176 ff.;
N. Lang, Laur. Aqu. ii, 165 ff.
189 D. Vuckovic-Todorovic, Starinar iV/15-16, 1966, 173 ff.
190 Arh. Pregl. v, 1963, II6 ff.
191 L. Vidman, SyllogeInscriptionumreligionis Isiacae et Serapiacae,ReligionsgeschichtlicheVersuche
und Vorarbeiten xxviii (Berlin, 1969), Nos 652-74, 7oo-7ooa.
192 Ibid., Nos 669, 670'
193 Cf. note 185.
194 For example,CIL xiii, 6646.
195 AI] pp. 144 if.
196 SeePannonia, 770 f.; 1\1S 222 ff., 234 ff.; a list of the Greekinscriptionsfound in Austria
is given by A. Betz, Wiener Studienlxxix, 1966, 593 ff.
197 Itinerarium Burdig. 566, 7.
19 8 Spomenikxcviii, 448, 449.
199 E.g. Starinar iV/7- 8, 293.
200 SHA, Aurelianus24, 3.
201 E.g. CIL vi, 2552, 2662, 32783.
202 G. Erdelyi, Arch. Ert. lxxvii, 1950, 72 ff.
203 C. 8098 = 6298,12659;Spomeniklxxi, 311.
204 For bibliography, seePannonia, 768 f.; recently G. Erdelyi, Acta Arch. xiii, 1961, 89 ff.;
Acta Ant. xiv, 1966, 211 if.; E. Diez, Carn.-]b. 1963-4 (1965),43 ff.; Actes du viiie Congres
Internat. d'Archiologie ClassiqueParis I96J (Paris, 1965), 209; Z. Kadar, ibid., 381 ff.; A. Sz.
Burger, Fol. Arch. xiii, 1961, 5I if.
205 E.g. Intercisa i, Nos 150-2, 165.
206 Ferri, ARD 226, Fig. 265.
207 C. 15166; cf. 4490; Pannonia, 768; Acta Arch. xxi, 1969,372.For a youth omnibusstudiis
liberalibus eruditusin Brigetio see Komarom Megyei MuzeumokKiizlembryeii, 1968, 81.
208 Gy. Di6sdi, Ant. Tan. viii, 1961, 99 ff.; MS 228 and 252.
209 C. 1°717,10864.
210 On these,seePannonia,769 f.; but cf. J. Gy. Szilagyi, Arch. Ert. xc, 1963, 189 if.; MS 232.
2I I Seethe summaryof the Vergilian passagesin R. P. Hoogma, Der Einjluss Vergils auf die
Carmina Latina Epigraphica(Amsterdam,1959)'It is evidentfrom this that Vergilian echoeson
the inscriptions of our provinceshave beentaken almost without exceptionfrom books i and
394
Notes to pp. 26}-p
vi of the Aeneid. On Vergil's influence on teaching,see,e.g., B. A. Gwynn, RomanEducation
from Cicero to Quintilian (Oxford, 1926), 154 if.
212 FragmentVat. 220.
213 MS 198-236.
214 E.g. Arch. Ert. 1944-5, Taf. xci; All 573; Schober,No. 206, etc.
215 E.g. the house with the Hercules-mosaicat Aquincum, 1. Wellner, Acta Arch. xxi,
1969, 271.
216 Aurelius Victor, Caes. 37,4; epit. 37,4; SHA, Probus 18, 8; 21, 2; Eutropius ix, 17, 2;
Eusebius,chron. 224a (Helm).
CHAPTER 8
I And for a short time to Florian also: Ann. Ep. 1970, 495.
2 Aurelius Victor, Caes. 40, 7.
3 C. 10275 = VHAD ix, 19°6-7,II2, No. 233.
4 Aurelius Victor, Caes. 39, II; SHA, Carinus 18, 2; ltinerarium Hieros. 564, 9.
5 SHA, Probus 18, 2; cf. Pannonia, 567; T. Nagy, Bp. Miieml. ii, 56.
6 SHA, Carinus 8, I; 9, 4; Eutropiusix, 18, I.
7 On the medallion of Numerianus,see G. Elmer, Der Miinzsammlerviii, 1935, 17 f.; cf.
Pannonia, 567.
8 Cf. Aurelius Victor, Caes. 39, 43: Diocletian as conquerorof the Marcomanni.
9 Cf. Aurelius Victor, Caes. 40, 9-10.
10 Eutropius,ix, 25, 2, speaksof ingentescaptivorumcopiae.
II Itinerarium Burdig. 565, 8.
12 Not. Or. xxviii, 26; Occ. xl, 54.
13 GentilesSarmatae:Not. Occ. xlii, 46-7°'
14 Th. Mommsen,GesammelteSchriftenii, 267 if.
15 Aurelius Victor, Caes. 39, 30 (Diocletian in Pannonia);W. Ensslin, PWREviiA, 2438 f.
16 A. Alfoldi, Arch. Ert. 1941, 5I if.; R. Egger,IOAI, xxxv, 1943,27.
17 Panegyrici Latini viii (Baehrens), 5, 1.
18 Codexlust. ix, 20, 10-II.
19 And perhapsin 303 on the journeyfrom Rometo Nicomedia:Lactantius,mort.pers.17,4.
20 Chron. Min. i, p. 230.
21 Mostly with PROVIDENTIA or VIRTVS, but at the mint of Rome also with VICTORIAE
SARMATICAE, RIC vi, p. 354, No. 43.
22 Not. Occ. xxxiii, 41; xxxiii, 48.
23 C. I0605a; cf. A. Alf61di, Arch. Ert. 1941, 49; R. Egger, IOAI, xxxv, 1943, 21 if.; W.
Ensslin, PWRE viiA, 2440. On the find-spot of the inscription, see A. M6csy in Limes-
kongress1969.
24 A. M6csy, Fol. Arch. x, 1958, 89 if. ; Pannonia,642; Og. Zb. xii, 1969,71 if.; Limeskongress
1969.
25 V. Balas, Acta Arch. xv, 1961, 310 if.; S. Soproni, Arch. Ert. xcvi, 1969,43if.
26 The Hungarian name for the earthwork is Ordogarok (Devil's Dyke) or Csorszarok
(Ditch of Csorsz).Csorszis a loan word from the Slav languageand also means'devil'. A date
can be derivedfrom this, since,when the Slavs arrived in the HungarianLowlands,the origin
395
Notes to pp. 272-9
of the earthworkwas no longer known. This gives a terminusante quemof the sixth or seventh
century.
27 S. Soproni, op. cit. (note 25).
28 P. Patay, Arch. Ert. xcvii, 1970, 312.
29 Chron. Min. i, 230; seenote 7.
30 Aurelius Victor, Caes. 40, 9-10.
31 Ammianus Marcellinus xxviii, I, 5; cf. Eutropiusix, 25, 2.
32 Aurelius Victor, Caes. 40, 10; Ammianus Marcellinus xix, I I, 4.
33 From Aurelius Victor, Caes.40, 10, one might concludethat the drainagewas begunonly
shortly before the death of Galerius, and as the naming of the province Valeria is connected
with it, the partition of PannoniaInferior had to be dated c. 310. But the partition must be
placed earlier than 303 accordingto C. 10981; cf. T. Nagy, Bp. Muem/., ii, 106, n. 367.
34 E.g. the omission of Savia in ItineraritJm Burdig., and the mention of tres Pannoniaeby
Optatusof Milevis, Corpus Scr. Ecel. Lat., xxvi, p. 33. Thesepiecesof evidencesuggestthat
Savia was not founded before the reign of Constantinethe Great.
35 Not. Occ. xxxii, 28, 40, 45 ; T. Nagy, Aktendesiv Internat. Kongresses fur griechischeund /atein-
ische Epigraphik, Wien, 1962 (Vienna, 1964), 274 if.
36 Not. Occ. i, 51, 83, 87; cf. Pannonia, 6Il. The praesesVa/eriae is missing in the Notitia;
cf. A. H. M. Jones,The Later RomanEmpire iii (Oxford, 1964), 351.
37 B. Saria, PWRExxi, 1174.
38 CIL vi, 32937.
39 Not. Or. i, 77, 121, 122, 124; cf. MS 41 if.
40 Eutropiusix, 25, 1.
41 Excerpta Va/esiana3.
42 For the date, see J. Moreau, Carn.-Jb. 1960 (1962),7if.
5
43 C. 4413.
44 Aurelius Victor, Caes. 40, 8.
45 Epi!. 41, 5; Eutropius x, 5; Excerpta Valesiana, 16-17; Zosimusii, 19, 1.
46 Z. Kadar, Pol. Arch. xii, 1960, 132 if.; Arheologija(Sofia) iii, 1961, 1,47if. I prefer this con-
text for some further hoards, e.g. the gold medallions and jewelry from Petrijanec: see J.
Kovrig, 'Die Haupttypen der KaiserzeitlichenFibeln', Diss. Pann. ii, 4 (Budapest, 1937),
PI. xxxviii.
47 Seeck,Regesten,162 if.
48 Zosimusii, 21; OptatianusPorphyr., carm. vi.
49 Codex Theod. i, I, 1.
50 An edict from Bononia, datedprobably to 6 July; Codex Theod. xi, 27, 2.
5I Chron. Min. i, 234; Eusebius,chron. 233C (Helm).
52 Excerpta Valesiana 32.
53 Excerpta Valesiana 32; Eusebius,chron. 233 f. (Helm); Eusebius,vita Const. 4, 6; Chron.
Min. i, 234.
54 Seenote 51.
5 5 Seenote 53·
56 Ammianus Marcellinus xvii, 12, 19.
5 7 Seeck,Regesten,I 82 f.
58 A. AlfOldi, Riv. Ital. di Numismaticaxxiv, 1921, II3 if.
396
Notes to pp. 279-87
59 Ammianus Marcellinus xvii, 13, 18-19.
60 Jordanes,Get. I 15, 161.
61 J. Korek and M. Parducz, Arch. Ert. 1946-8, 298 if.; I. Bona, in Oroshdza Tb"rtenete i
(Oroshaza,1965), 114 if.
63 D. Tudor, Oltenia Romana(third ed., Bucharest,1968), 25 I if.
64 Tokod: Acta Arch. xxi, 1969, 35 I, Abb. 3. Esztergom-Hidegleloskereszt: S. Soproni,
LRKN 137, Abb. 4. Pilimarot: ibid. 136, Abb. 3. Visegrad-Sibrik:ibid. 134, Abb. 2. Slanka-
men: D. Dimitrijevic, Starinar ivl7-8, 1956-7, 304; Rad. vojv. muzo v, 1956, 150 if.
65 For exampleBudapest-Eskliter: L. Nagy, AZ Eskii teri rrfmai erOd, Pestvdrosiise (Budapest,
1946). For Visegrad-Sibrik,seenote 64.
66 Lebeny-Baratfoldpustzta: D. Gabler,Arrabonaviii, 1966, 67 if.; Arch. Ert. xcv, 1968, 131.
Almasflizito: F. Flilep, Arch. Ert. lxxxvii, 1960, 236. Szentendre:T. Nagy, Arch. Ert. 1942,
262,27°.Nagyteteny:F. Flilep and E. Cserey,Nagytetef!Ymiiemlekei(Budapest,1957).Intercisa:
L. Barkoczi, Intercisa i, 32 f. IZa oppositeBrigetio: B. Swoboda,Slov. Arch. x, 1962,422if.
67 C. 15172; Not. Occ. xxxiii, 34.
68 Nagyteteny (Campona), Intercisa and perhaps Szentendre (Ulcisia Castra= Castra
Constantia).
69 See, for example,Tokod, Nagyteteny,Intercisa.
70 E. Nowotny, RUG xii, 1914.
71 Provedbest at Tokod (unpublished),where all the buildings are againstthe fort-walls.
72 See,for example,Tokod, Pilismarot.
73 Tokod, Budapest-Eskliter, Visegrad, Szentendre.
74 Cortanovci, seeM. Manojlovic, Radvojv. muzo xi, 1962, 123 if.
75 Stampswith namesof placesandstampsof legionesVII CI. and IIII FI. are limited to the
limes of Moesia Prima, stampswith DRP to Dacia Ripensis.
76 D. Tudor, SCIVxi, 1960, 335 if.
77 A. Alfoldi, Arch. Ert. 1941, 53.
78 ILS 724.
79 Seeck,Regesten,187 if.
80 Ibid., 186 f.
81 Ibid., 188 f., 191, 195.
82 Aurelius Victor, Caes. 41, 26; Eutropiusx, 10,2; Eusebius,chron. 237C (Helm).
83 Zosimusii, 44, 3; Julian, or. ii, 76d; Eusebius,chron. 238c (Helm), etc.; cf. Seeck, Ge-
schichtedes Untergangsder antiken Welt iv, 429 f.
84 This villa (AmmianusMarcellinus xxvi, 5, I) has beenpartially excavated,see Chapter 9
(p. 302).
85 Chron. Min. i, 238; Eusebius,chron. 238e (Helm).
86 Cf. A. Mocsy, Ant. Tan. xiii, 1966, 242 if.
87 Zosimusii, 46-5°.
88 Ammianus Marcellinus xvi, 10, 20; Zosimusiii, I, 1.
89 A coin-hoardin the north-eastof Pannoniamay also be referred to it: K. B. Sey, Fol.
Arch. xvi, 1964, 63 if.
91 Zosimusiii, 2, 2; Julian, ep. ad Athen. 279d.
397
Notes to pp. 287-98
92 Seeck,Regesten,204 ff.
93 xvi, 10; xvii, 12-13; xix, II.
94 xvii, 13, I.
95 Seeck,Regesten,206 f.
96 Ammianus Marcellinus xxi, 9-10; Zosimusiii, 10-11.
97 Ammianus Marcellinus xxi, 12, 21-3; xxii, 2, 2.
98 For exampleC. I0648b = ILCV, II.
99 Ammianus Marcellinus xxx, 5. On the non-existenceof a Pannonianparty under Valen-
tinian, seeA. Demandt,Historia xviii, 1969,618.
100 Epi!. 44, I.
101 Ammianus Marcellinus xxx, 7, 2; Eusebius,chron. 243e (Helm), etc.
102 Ammianus Marcellinus xxvi, 5, I.
103 Seeck,Regesten,215 ff.
104 Ammianus Marcellinus xxvi, 4, 5.
105 Mosella, 8-9; cf. Not. Occ. xlii, 65-70.
106 See,for example,Ammianus Marcellinus xxvi, 4-5.
107 Codex Theod. xv, I, 3.
108 Pannonia, 629 ff.; Eirene iv, 1965, 143 f.
109 S. Soproni, Stud. Mil. 138 ff.
110 C. 3653; 10596; Soproni, Stud. Mil. 138.
III Seethe works of S. Soproni, especiallyStud. Mil. 138 ff., and LRKN 131 ff.
II2 C. I4358II.
I I 3 For burgi, seePannonia, 638 ff.; S. Soproni, op. cit. (note III), StudiaComitatensiai, 1972,
39 ff., and D. Dimitrijevic, o.j. Zb. xii, 1969, 88 ff. A large number of burgi excavatedby S.
Sopronihavenot beenpublished-see so far Eireneiv, 1965, 141 f.; Acta Arch. xxi, 1969, 354 f.
114 xxix, 6.
I I 5 S. Soproni, Ig6g Congressof RomanFrontier StudiesReport(Cardiff, 1974).
116 Not. Occ. xxxii, 41; xxxiii, 48.

117 xxxix, 6, I.
I 18 In ipsisQuadorumterris, quasi Romanoiuri iam vindicatis.
119 M. R. Alf6ldi, AntiquitasHungarica iii, 1949, 86 ff.; A. Sz. Burger, Num. KiiZI.lxvi-lxvii,
1967-8,3ff.; cf. M. R. Alf6ldi, Jahrbuchfur Numismatikund Geldgeschichtexiii, 1963, 93 f., andin
addition A. Mocsy, Eirene iv, 1965, 141 f.
120 Cf. V. Lanyi, Acta Arch. xxi, 1969, 33 ff.; K. B. Sey, Fol. Arch. xx, 1969,63ff. They may
belong to the barbarianraids of 379.
121 Ammianus Marcellinus xxx, 5-6.
122 SeeA. Mocsy, in Adriatica, Melangesofferts d Grga Novak (Zagreb, 1970), 583 ff.
123 Ammianus Marcellinus xxx, 10; Chron. Min. i, 242; Socratesiv, 31, 6-7.
CHAPTER 9
I See, for example, R. Noll, Barb-Festschr. 160 f., where the late Romaninscriptions of
Austria are listed.
2 Solinusxxi, 2; Expositio tot. mundi, 57; Avienus, descr. orbis, 456-7; Isidore, orig. xiv, 4, 16.
3 Itinerarium Burdig. 563, 10; cf. Tabula Imperii RomaniL 34 (Budapest,1968), 59.
398
Notes to pp. 298-}02
4 SeeChapter7, note 216 (p. 395)·
5 C. 10275 = VHAD ix, 1906-7, IIZ, No. 233.
6 Itinerarium Ant. 243; Not. Occ. xxxii, 45.
7 Pannonia, 669, and 1. Vincze, Acta EthnographicaAcad. Sc. Hungaricae vii, 1958.
8 B. Saria, Der riimische GutshofvonWindenam See(Eisenstadt,1951), 16 fr.; Bp. Reg. xv, 1950,
313; Arch. Brt.lxxviii, 1951, 128 f.
9 xlix, 36. 2.
10 Herodianviii, 2.
II Ambrose,ep. xviii, 21; cf. S. Panciera,La vita economicadi Aquileia (Venice, 1957), IIO f.;
L. Ruggini, L'Italia Annonaria (Milan, 1961), II3 fr., 534.
12 Seenote 2.
13 Thomas,R VP 128 fr.
14 Cf. A. M6csy, Eirene iv, 1965, 149 f.; Acta Arch. xxi, 1969, 365.
15 Seeespeciallythe bath-buildingsnorth of Lake Pelsoin Thomas,R VP: Balatonfokajar:
21 f.; Balatonfiired: 23 f.; Balatongyorok: 25 f.; Rezi: III, and Abb. 5 on page 21. Nearly
all were uncoveredin old excavationsand not dated.
16 See previous note; also Orvenyes: T. Szentle1eky,VMMK iv, 1965, 103 fr.; Kekkut:
K. Sagi, Arch. Brt. xciii, 1966, 295; Siimeg: Thomas,RVP IIZ; also part of Szentkiralysza-
badja: Thomas,R VP II 8 fr.; etc. South of Scarbantia:D. Gabler, ArchiiologischeForschungen,
Mitteilungen des ArchiiologischenInstitutesder UngarischenAkademieii, 1971, 57 fr.
17 Pecs-Mecsekalja:Thomas, RVP 288; Hosszuheteny:Thomas, RVP 274; Kekkut:
K. Sagi, op. cit. (note 16); cf. Thomas,R VP 55; Siimeg: Thomas,R VP I 12; Tac I: J. Fitz,
Gorsium (Szekesfehervar,1970).
18 Gyulafirat6t III: Thomas, RVP 44. Siimeg: Thomas, RVP II2. SzentkiraIyszabadja:
Thomas, RVP II8 f. Hosszuheteny:Thomas, RVP 274f. Koml6: see Chapter7, note 108
(p. 391); samianand early Romanfinds are not known at this villa.
19 Siimeg, Pecs-Mecsekalja,etc.
20 For example,Tiiskevar: A. Kiss, VMMK vi, 1967, 37 fr. Purbach:Thomas,RVP 192 fr.
Egregy: Thomas,RVP 33, etc.; seenote 15.
21 Csucshegy:L. Nagy, Bp. Reg.xii, 1937, 25 fr. and J. Szilagyi, Arch. Brt.lxxvi, 1949,67fr.;
Budakalasz:Thomas,R VP 214 f., etc.; domesticbuildings in the environsofVindobonaalso:
A. R. Neumann,Barb-FestschriftII 5 fr.
22 ILS 8987; for dating, seeActa Arch. xxiii, 1971, 357.
23 D. Sabovljevic,Starinar v, 1888,66fr.
24 M. Dimic, Starinar viii, 1891, 21 fr.
25 M. VelickoviC, ZNMB i, 1958, IIO fr.
26 M. Vasic, RevueArcheologique1903, 19 fr.; Garasanin,Nalazilta, Taf. xxivb, etc.
27 M. Grbic, Plastika, Taf. lxviii.
29 Epit. 40, 10.
30 Epit·40 , 16; Procopius,aedif. iv, 4, 3.
3I Ammianus Marcellinus xxvi, 5, I.
32 A. OrSic-Slavetic, Starinar iii/8-9, 1933, 307 fr.; iii/13, 1938, 199 fr.; Spomenikxcviii,
233-48; cf. MS 91 f.; A. Nenadovic,LuJi, 1961, 169.
33 B. Saria, Barb-Festschrift25Z fr.
399
Notes topp. }02-I2
34 xiv, II, 20; S. Jenny, Mittheilungen der Zentral-Commissionxxii, 1896, Iff.
35 xxx, 5·
36 D. Mano-Zisi, ZNMB ii, 1960, 100 ff.; Arch. Iugosl. ii, 1956, 72 ff.; La mosai"quegreco-
romaine (Paris, 1965), etc.
37 Pannonia, 700 f.; K. Sagi, Acta Arch. ix, 1961, 397 ff.
38 Most recently L. Barkoczi, Acta Arch. xx, 1968,275 ff.; K. Sagi, Acta Ant. xviii, 1970,
147·
39 K. Sagi, Acta Arch. xii, 1960, 190 ff.
40 MS III ff.
41 S. Soproni, Arch. Ert. xcvii, 1970, 310; xcviii, 1971, 271.
42 A. Radnoti, PWREviiA, 82 f.
43 A. Sz. Burger, Acta Arch. xviii, 1966, 99 ff.
44 A. Radnoti, Laur. Aqu. ii, 91 ff.
45 E. Biro, Arch. Ert. lxxxvi, 1959, 173.
46 Mursella: Codex Theod. xvi, 8, I; cf. i, 8, 6 and 9, 2 (A.D. 339)' Tricciana: Codex Theod.
xi, 36, 26 (A.D. 379), doubtful, see Seeck,Regesten109.
47 C. 421 9.
48 Not. Occ. xii, 2I.
49 L. Barkoczi, Acta Arch. xx, 1968, 275 ff.; A. Kiss, Arch. Ert. xcv, 1968, 93 ff.
50 MS 67,72.
5I For exampleC. 3522, 8I 5I; cf. p. 1022.
52 C. 4180.
53 For example C. 3370, 3576,4°39,4219,10527, 10981, 13810, 14594, 15172; Intercisa i,
No. 32; Spomenikxcviii, 229, etc.
H On the discontinuationof the practice of setting up inscriptions, see MS 203. The sug-
gestionsmadethere are not fully valid for Pannonia;cf. also A. Mocsy in G. AlfOldy, Gesell-
schaftund Bevolkerungder ro'mischenProvinz Dalmatien (Budapest,1965), 214.
55 Panegyrici Latini ii (Baehrens),37.
56 Codex Theod. i, 32 , 5.
57 L. Nagy, Bp. Reg. xii, 1937, 25 ff.; cf. K. Sz. Poczy, Bp. Reg. xxii, 1971,98ff.
58 L. Nagy, Mumienbegriibnisse; Bp. Reg. xiv, 1945, 535 ff.; Gy. Parragi, Bp. Reg. xx,
1963, 3II ff.; K. Sz. Poczy, Arch. Ert. xci, 1964, 175, etc.
59 K. Sz. Poczy, Bp. Reg. xxi, 1964,62ff.; T. Nagy, Bp. Miieml. ii, 59 ff.
60 Swoboda,Carnuntum 140 f.
61 xxx, 5-6.
62 See, e.g., Chapter 7, note 214 (p. 395); in addition jOAIB xxxi, 101 ff.; cf. MS 105 f.,
109, 121.
63 Seenote 5I; also C. 1661,4°39,4413,10981, 12657,etc.
64 MS 105, 109, 122; M. Velickovic, ZNMB iii, 1962, 99 ff.
65 K. Sz. Poczy, Arch. Ert. xciv, 1967, 137 ff.
66 Antiquiry x (1936), 475. In the caseof Bassianaa later date for the town-wall cannot be
excluded,seeChapter10 (p. 348).
67 D. Boskovic,Starinariii/4, 1928, 270, Abb. I; B. Saria,BRGK xvi, 1925-6, 93; D. Piletic,
Arh. Pregl. iv, 1962, 176 ff.; D. Mano-Zisi and Lj. Popovic, Starinar iV/9-10, 1959, 381 ff.
68 For exampleAlj, p. 269 (Aquae Balisae); J. Korda, Luj i, 1961, 59 ff. (Cibalae) etc.; cf.
400
Notes to pp. }I2-20
Pannonia, 696. The problemof Velike Malence,the large mountainfortress nearNeviodunum

(Fig. 50), remainsunsolved.It may have servedas a refuge for the town. For the most recent
plan seeP. Petru, Olj. Zb. xii, 1969, Fig. 14.
69 C. 10107 = ILS 3458.
70 Seethe works cited in Chapter 5, note 9 (p. 378).
71 Seenote 70.
72 A. Hytrek, EphemerisSalonitana(Zara, 1894), 5 if.; V. Hoffiller, Bericht iiber den vi Internat.
Kongressfiir Archiiologie (Berlin, 1939), 522 f.
73 Grbie, P lastika, 127 if.
74 Starinarv, 1888, Taf. x; iii/8-9, 1933, 304 if.; iV/5-6, 1956,53 if.; Arch. Iugosl. ii, 1956,
85 if.; Starinar ivl9-10, 1959, 382 if.; iV/II, 1961,246if.; LuJ i, 1961, 171 if.; Jahrbuch des
Riimisch-Germanischen Zentralmuseumsx, 1963, 118 if., etc.
75 L. Mirkovie, Arch. Iugosl. ii, 1956, 85 if.
76 L. Nagy, RiimischeMitteilungen xli, 1926, 123 if.; Fr Gerke, in Neue Beitrage zur Kunstge-
schichtedes i. Jahrtausends(Baden, 1954), 147 if.
77 Gy. Gosztonyi,A pecsi okeresztenytemetij (Pees,no date); F. Flilep, Acta Arch. xi, 1959,
401 if.; Arch. Ert.lxxxix, 1962, 23 if.; xcvi, 1969, 3 if.
78 1. Paulovics,Szombathe!JSzentMdrton egyhdzdnakromaikori eredete(Szombathely,1944).
79 The publicationof the excavationsby 1. Paulovics,Savaria-Szombathe!J topogrdftcija (Szom-
bathely, 1942),is in somerespectsnot satisfactory.A new compilationof the evidencefrom the
excavation,basedon photographsand notes,has beenmadeby my studentE. Toth, who has
solved the problem of the aula.
80 Not. Or. ix, 44. For the cemetery,seeGMKM i, 1956, 320 if.; ii, 1957, Abb. 3; iv-v, 1960,
371 if.; D. Mano-Zisi, La mosai'quegreco-romaine(Paris, 1965), Fig. 18.
81 Starinar iV/7-8, 1957, 289 if.; tA vi, 1956, 292; xiii-xiv, 1964, 148 if., etc.
82 Moravski arheoloIki Glasnik iii, 1936, 43; seealso MS 92 f.
83 M. Valtrovie, Starinar i, 1884, 62 if., 90 if.; M. Vasie, Jahrb. DAI xx, 19°5; Archiiol.
Anzeiger, 108; N. Vulie, ibid. xxvii, 1912, 548.
84 The best-knownis the discovery of the sarcophagusat Szekszard:L. Nagy, PS 48 if.
85 Not. Or. ix, 40, 43, 44·
86 On Tac = Gorsium, see most recently J. Fitz, Acta Arch. xxiv, 1972, 3 if. Matrica: A.
M6csy, Arch. Ert. lxxxii, 1955,62if.
87 Cf. Eirene iv, 1965, 147 f.; Acta Arch. xxi, 1969, 365.
88 E. Nowotny, RUG xii (1914),
89 Unpublished.
90 Seenotes 26, 27, 84, 141, 144, and 145·
91 For a comprehensivestudy of the late Roman cemeteriesof Pannonia,see V. Lanyi,
Acta Arch. xxiv, 1972, 53 if.
92 Seenote 91.
93 Pannonia,765 with bibliography; seealso Z. Kadar, Arch. Ert. xcv, 1968,90if.; Fol. Arch.
xv, 1963, 69 if.; B. Jelick, ZNMB iii, 1963, 109 if.; and D. Gaspar, Spiitrijmische Kiistchen-
beschliigein Pannonien(Szeged,1971).
94 L. Nagy, Arch. Ert. xliv, 1930, III if.; Bp. Reg. xii, 1937, 189 if.; L. Barkoczi, Fol. Arch.
xii, 196o, 121 if.
95 M. R. Alf6ldi, Intercisa ii, 404 f.
401
Notes to pp. j2I-j2
96 There is no study of coin-circulation for Pannoniaand Upper Moesia; see Pannonia,
691 ff.; MS 2.57 ff.
97 E.g. L. Nagy, MumienbegrlibnisseIS f.; K. Sz. Poczy, Arch. Erl. xci, 1964, 176 ff.; L. Bar-
koczi, Acta Ant. xiii, 1965, 2.51.
9S This is proved by finds of hoards, see, for example, Num. KoZI. xxviii-xxix, 192.9-30,
30 ff.; Arch. Ert. xciii, 1966, 2.96. On captivesseee.g. OptatianusPorphyrius,carm. vi.
99 C. 3653.
100 I. Paulovics,Arch. Ert. xlvii, 1934, 163.
101 E.g. the martyrs of Sirmium: Demetrius,Basilla, Anastasia,Synerotas,Hermagoras,etc.
102. The earliest-knownbishopsare Eusebius,Irenaeus,Quirinus, Domnus and Victorinus,
but for the latter seenote 103.
103 Jerome,vir. ill. 74.
104 For Syriansin Sirmium in post-Severantimes, seeC. 2.006 and 6443.
105 The Christiantombstonesare difficult to date. Thosefrom Sirmium are not completely
listed in ClL iii; for these,see VHAD x, 19o5-9, Nos. 374-466(pp. 2.31-63).
106 C. 4413; seealso C. 4039 (Poetovio).
107 See,for example,~the objectsfrom the Mithraeumat Sarkeszi:T. Nagy, Bp. Reg.xv, 1950,
50.
lOS S. Soproni,Arch. Ert.lxxxi, 1954, 50; Spomeniklxxv, 171 = Starinar iii/S-9, 1933, ;ro,
No. II.
109 M. Abramic, Fuhrer durch Poetovio (Vienna, 192.5), 6S; B. Saria, Starinar iii/3, 192.4-5,
163 = BRGK xvi, 192.5-6, 92..
lIO M. K. Kubinyi, Arch. Ert. 1946-S, 2.76 ff.
III Fr Kenner, Mitteilungen der Zentral-Commissionxii, IS67, 12.7 ff.
II 2. For example,Soproni, Szemlex, 1956, 346.
II3 T. Nagy, Bp. Tort. i, 446 n. 2.15; C. Praschniker,IOAl xxx, 1937, 12.0 f.
114 Pannonia, 749 f.
I I 5 A. Alf6ldi, Arch. Ert. 1940,2.14ff. on C. 3343; but cf. R. Eggerin Omagiului C. Daicovi-
ciu (Bucharest,1960), 167 f., and Pannonia, 595.
II6 L. Barkoczi, Acta Ant. xiii, 1965, 2.3S ff.
I I 7 E.g. Passio Irenaei 3-4.
lIS Acta Sanctortlm,Aprilis, iii, p. 571.
II9 Corpus Scr. Ecel. Lat. xlix.
12.0 Cf. Passiolrenaei 5.
12.1 Seemost recently D. Simonyi, Acta Ant. viii, 1960, 165 ff.
12.2. For the following, seeJ. Zeiller, Lesorigines chretiennesdanslesprovincesDantlbiennes(Paris,
19I5); T. Nagy, 'A pannoniai keresztenysegt6rtenete',Diss. Pann., ii, 12. (Budapest,1939);
R. Egger, Der Heilige Hermagoras(Klagenfurt, I94S).
12.3 For exampleC. 102.32, 102.33.
12.4 Priscus,fragment 2..
12.5 Egger, RAFCi, 57 ff.
12.6 For exampleR. Bratanic, Arh. Vestn. iv, 1953, 2.S2. ff.; also VHAD viii, 1905, No. IS7;
ix, 1906-7, Nos 190-1; IOAlB, iii, 134, No. 37; vi, 2.1, Nos 2.S-9; xii, 15S, No. 2.3; xiii, 2.04,
Nos 15-16.
12.7 Seenote 79.
402
Notes to pp. jjj-4I
128 I. Paulovics,SzombathefySzentMdrton egyhdzdndkrrfmaikori eredete(Szombathely,1944),
31, Abb. 10.
129 Seenote 123.
130 L. Nagy, AZ rfbudai rfkeresztitrycella trichora a Raktdr-utcdban(Budapest,19F); V. Roffiller,
Bericht uber den vi Internat. Kongressfur Archiiologie (Berlin, 1939), 522 f.; F. Flilep, Acta Arch.
xi, 1959, 401 ff.
IF SeePannonia, 727 f.; F. Flilep, Arch. Ert. lxxxix, 1962,23ff.; xcvi, 1969, 3 ff.
IF A. Radn6ti, Arch. Ert. Iii, 1939, 152 ff.; E. Cerskov, GMKM iv-v, 1960, 374, Abb. 3.
133 Starinar viii, 1891, 130 ff.; iii/8-9, 1933,75,Abb. 3.
134 L. Nagy, PS 48 ff.
135 SeeL. Mirkovic, Arch. Iugosl. ii, 1956, 85 ff.
136 F. Flilep, Arch. Ert. xcvi, 1969,33.
137 Ibid., 3 ff.
138 Gy. Gosztonyi, Arch. Ert. 1940, 56 ff.
139 Ulpianum: GMKM i, 1956, F2; vii-viii, 1964, 352 ff. Remesiana:Starinar iV/9-IO,
1959, 381 ff. Aquincum: Arch. Ert. 1940, 250 ff. Fenekpuszta:Acta Ant. ix, 1961, 397 ff.
Kekkut north of Lake Pelso: Arch. Erl. xlv, 1931, 32 ff.; but cf. K. Sagi, Arch. Erl. xciii,
1966, 295; seealso Pannonia, 755 f., and Acta Arch. xxi, 1969, 370.
140 C. Praschnikerin Kubitschek, Ri;,merfunde52 ff.; cf. A. Barb, Mullus, Festschriftfur Th.
Klauser (Mlinster, 1964), 17 ff.; E. B. Thomas,Acta Ant. iii, 1955, 261 ff.; Orvenyes:unpub-
lished, in the Lapidarium of the museumat Tihany.
141 A. Rekler, Arch. Ert. xxxiii, 1913, 210 If., cf. G. Erdelyi, Ant. Tan. viii, 1961, 137 f.
142 Grbic, Plastika, Taf. lxviii.
143 A. AlfOldi, Acta Arch. (Copenhagen)i, 1934, 99 ff.; M. Manojlovic-Marijanski,
Kasnorimskifljemovi iZ Berkasova(Novi Sad, 1964).
144 See Chapter8, note 46 (p. 396); also M. Lenkei, Fol. Arch. vii, 1955,97If. and note 26
above.
145 G. Erdelyi, Arch. Ert. xlv, 193 I, I If.
146 Seenote 93.
CHAPTER 10
Jerome,ep. 60, 16, 2 = MPL xxii, 600 = ed. Labourt, p. 106 f.; cf. ep. 123, 16 = MPL
xxii, 1058 = ed. Labourt, 92 f.
2 Chron. Min. ii, 76.
3 Seeck,Regesten,250 f.
4 The main sourcefor theseeventsis Ammianus Marcellinus xxxi, 4-16. Seenow T. Nagy,
Acta Ant. xix, 1971, 299 ff.
5 Chron. Min. i, 243, 297; ii, 60.
6 In July he was still in Scupi: Codex Theod. vi, 39, 2; two edicts dated to August are from
placesas yet unidentified: Codex Theod. vi, 30, 3; xii, 13,4.
7 SidoniusApollinaris v, 107 If.; for his rank, seeVarady, LJP 38.
8 Chron. Min. i, 243.
9 Jordanes,Get. 139-42; Zosimusiv, 31.
10 Vitalianus: Ammianus Marcellinus xxi, 10, 9; Zosimus iv, 34. For destructionby the
403
Notesto pp. 34I-J
peopleof Alatheusand Saphracin Pannonia,seePanegyriciLatini ii (Baehrens)32, 3-4; Colleetio
Avellana 39, 4 (Corpus Ser. Eeel. Lat. xxxv, 89); Jerome,vir. ill. 65; Comm. in Soph. i, 676
(MPL xxv, 1340 f.); cf. Chapter8, note 120 (p. 398).
II Zosimusiv, 32-3.
12 PanegyrieiLatini ii (Baehrens)32, 3-4; Jordanes,Get. 139-42; for the emperors'meeting:
Seeck,Regesten,254 f.; cf. Varady, LJP 378. For the Gothic treaty of 382, see,e.g., Chron. Min.
i, 243; ii, 61; Panegyrici Latini ii (Baehrens)22, 3; Themistius,or. xvi, 208.
13 CILv, 1623; cf. Egger, RAFCi, 57ff.
14 Cassiodorus,varia v, 14,6; cf. Varady, LJP 522.
15 xxxi, 4, 2.
16 Num. KijZI. xxviii-xxix, 1929-30, 30 ff.; Arch. Ert. xciii, 1966, 296.
17 Ambrose gives a concise summary of the unrest among the barbarians:in Luc. x, 10
(= MPL xv, 1898 f. = Corpus Scr. Ece!. Lat. xxxii, 458 f.).
18 Symmachus,Relatio 47.
19 See Chapter9, note I I (p. 399).
20 ILS 8987; for the date, seeActa Arch. xxiii, 1971, 357.
21 Zosimus iv, 42, 5.
22 Zosimus iv, 45, 3; 48-5 c.
23 Panegyriei Latini ii (Baehrens)34; 37; Seeck,Regesten275.
24 A. Alf6ldi, Der Untergangder Rijmerherrschaftin Pannonie11i (Berlin-Leipzig, 1924), 13.
25 Ibid. 15.
26 V. Lanyi, Acta Arch. xxi, 1969, 33 ff.
27 A. Alf6ldi, EgyetemesPhilologiai KO'zlijny !iv, 1930,2ff.; L. Bark6czi, Fol. Arch. xiii, 1961,
I I I f.; S. Soproni, Fol. Arch. xx, 1969, 69 ff.
28 Zosimusiv, 48-9'
29 Ambrose, de obittf Valent. 2, 4, 22 (MPL xvi, 1427 ff.).
30 Vir. ill. 65.
31 Comm. in Soph. i, 676 (MPL xxv, 1340 f.).
32 Ep. 66, 14 (MPL xxii, 647 = ed. Labourt, 180).
33 For example,Jerome,ep. 60, 16, 2; Comm. in Soph. I, 676; Ambrose, de off. ministr. i, 15,
70 (MPL xvi, 129); in Luc. x, 10; dejide, ii, 140 (MPL xvi, 613); Ps.-Augustine,quaest.vet. et
nov. test. 115,46 (Corpus Scr. Eccl. Lat. i, 334); Collectio Avellana, 38, I (Corpus Scr. Ecel. Lat.
xxxv, 85); see also Claudian, bell. Goth. 632 ff.; in Rttj. ii, 26-53; Orosius vii, 43, 4. Varady,
LJP 123 ff., wants to reject theseand other points as literary platitudes.
34 Sidonius Apollinaris, vii, 589 f.; Ennodius,v. Ant. 12-14.
35 Cons. iii Std. ii, 191 fr.
36 L. Varady describesthe part played in the political and military history of this period by
thefoederatisettledin Illyricum (LJP). I cannotaccepthis commentson conditionsin Pannonia
at this time; cf. Acta Arch. xxiii, 1971, 347 ff., and T. Nagy, Acta Ant. xix, 1971, 299 ff.
37 The evidencefor the time and placeof the barbarianraids is given by Claudian,in Ruf. ii,
124; seealso ibid. I, 301-22; ii, 26 ff.
38 E. Polaschek,NumismatischeZeitschrift lviii, 1925, 127 ff.
39 Paulinus,vita Ambrosii, 36 (MPL xiv, 42).
40 Not. Occ. xxxiv, 24.
41 Jerome,ep. 60, 16, 2; 123, 15; cf. alreadyA mmianus Marcellinus xxxi, 4,2.
404
Notes to pp. 345-5}
42 E.g. in Ruf. i, ,10; epith. Pall. 88;fesc.iv, 15.
4, Cons. Stil. iii, 13.
44 ii, 191-207.
45 Cf. Claudian, bell. Goth. 279, ,6,-5.
46 Claudian,vi cons. Hon. 227 f.; Zosimusv, 26, 29; Jordanes,Get. 147.
47 Zosimus v, 37.
48 There is no evidencefor the route of this march and this, in my opinion, shows that the
Pannonianterritory through which the peopleof Radagaisustravelled had passedcompletely
outsidethe sphereof interestof contemporaries.So whetherthe march followed the Danube
valley or was partly outsideRomanterritory is irrelevant.
49 Ep. 12" 15 (MPL xxii, 1057).
50 Appendix, passioQuirini (Ruinart, Acta marryrum sincera 524); cf. T. Nagy, Regnumvi,
1947,244ff.; Egger,Der Heilige Hermagoras(Klagenfurt, 1948), 5If. T. Nagy, op. cit. (note ,6),
doesnot excludethe possibility that the body of Amantius camethroughtranslatio to Aquileia.
5I CodexTheod.x, 10, 25 ; cf. v, 7, 2. A younglady of senatorialrank went to Salona:C. 9515,
cf. ]RS lvii, 1967,299,and]. Wilkes, Phoenix 26, 1972, ,77 ff.
52 A. Chastagnol,Epigraphicaxxix,1967, 105 ff.; CodexTheod.xi, 17,4; xv, 1,49.
5, Zosimus v, 46, 2.
54 The omission of two frontier forts, Ad Mures and Statuas, between Arrabona and
Brigetio is equally notable. They were either the most easterlyforts of PannoniaPrima or
the mostwesterlyof Valeria, but they occurneitherin Occ. xxxiii nor in Occ. xxxiv. From this it
may be deducedthat theseforts no longer had garrisonswhen the chapteron PannoniaPrima
was broughtup to datefor the last time. If they had belongedto Valeria they would havebeen
included in the relevant chapterxxxiii, especiallyas this chapterdescribesan earlier, or even
severalearlier situations.
55 J. Zeiller, Les origines chretiennesdans les provincesDanubiennes(Paris, 1918), 148 ff.
56 Innocentiipapaeep. 42.
58 Priscus,fragment 2.
59 Ibid., fragment 7·
60 Cassiodorus,var. xi, I, 9; cf. Varady, L]P ,08.
62 Priscus,fragment 7.
6, Procopius,aedif. iv, 5.
64 Sidonius Apollinaris vii, 589 ff.
65 Aedif. iv, 5, 9·
66 Spomeniklxxvii,,8-9;](jAIxxxi, I I 7, No. 24; cf. also](jAIvi, 57, No. 90; viii, 6, No. 15,
etc.
67 Seenote 55.
68 Seenote 50.
69 Seenote 51.
70 Admittedly this hasnot beenrecordedas a flight, but as bishopof a Pannoniancommunity
he could not otherwisehave spent his last years in his Italian home. Seealso note 50.
71 On the cult of Pannonianmartyrs outsidePannonia,seeT. Nagy, A pannoniai kereszte~yseg
tiJ'rtenete 68 ff.
405
Notesto pp. 353-8
72 K. Sagi, Acta Ant. ix, 1961, 397 f.
73 L. Barkoczi, Acta Arch. xx, 1968, 275 ff. On Romanstransferredby the Avars from the
Balkansand from Italy to Pannonia,see1. Bona, Arch. Ert. xcvii, 1970, 258.
74 Ennodius(Monum. Germ. Hist. Auct. Ant. vii), vita Antoni 12-13.
75 On Leonianus and other refugees,see A. Alfoldi, 'Tracce del cristianesimonell'epoca
delle grandi migrazioni in Ungheria',Quadernidell'Impero: Romae Ie province (Rome, 1938).
76 Gregory of Tours, Hist. Francorum v, 37; Isidore, vir. illustr. xxxv; cf. J. Vives, Inscrip-
cionescristianas de la EspanaRomanayVisigoda (Barcelona,1942), 82 f., No. 275.
77 PaulusDiaconusii, 26.
78 C. 9551 = ILCV 1653; C. 9576 = ILVC 445 5; ILCV II8.
80 Procopius,aedif. iv, 5.
8I The father of Romulus, Orestes:Priscus,fragment 7; Excerpta Valesianaii, 38.
82 Priscus,fragment 7; and cf. ibid., to mention only the following: Rusticius, Constantio-
Ius, the builder of the baths of Onegesius,etc.
83 The very plausiblehypothesisof 1. Bona, Ant. Tan. xvi, 1969,285ff.
84 1. Bona, 'Die pannonischenGrundlagen der langobardischenKultur', Problemi della
civilta ed economialangobarda(Milan, 1964). For stonebuildings of post-Romandateat Intercisa,
seeE. B. Vago, Alba Regiaxi, 1971, II2.
85 Comm. in Isaiam vii, 19 (292).
86 D. Rendic-Miocevic,StarohrvatskaProso/etaiii/I, 1949, 9 ff.
87 V. Besevliev,EtudesBalkaniquesi, 1964, 147 ff.
88 Procopius,anecd. 6, 1.
89 For example,Jordanes,Get. 265, 282.
406
Select Bibliography
GENERAL WORKS AND WORKS OF REFERENCE
es
A. ALFOLD1, Magyarorszdgnepei a rdmai birodalom (Budapest,1934).
A. ALFOLD1, 'Pannoniaromaisaganakkialakulasa es t6rteneti kerete', Szdzadoklxx, 1936,
1. ff., 129 ff.
A. ALFOLD1, 'Studi ungheresisulla romanizzazionedella Pannonia', Gli Studi Romani nel
mondoii, 1935, 265 ff.
A. BETZ, Aus Osterreichsriimischer Vergangenheit(Vienna, 1956).
S. BORZSAK, 'Die Kenntnisse des Altertums tiber das Karpatenbecken',Diss. Pann. i, 6
(Budapest,1936).
Bp. Tort.
M. FLUSS, 'Moesia', PWRExv,1932, 2354ff.
B. GEROV, 'Romanizmatmezdu Dunava i Balkana' i-iii, Godilnik na Soft.jskija Universitet,
Filologileski Fakultet xlv, 1950; xlvii, 1951; xlviii, 1955.
T. KOLN1K, 'Prehl'ada staY badaniav dobe rimskej a stahovanenarodov', Slov. Arch. xix,
1971, 499 ff.

M. M1RKOV1C, Rimskigradovi na Dunavu (Belgrade,1968).
A. MOCSY, Gesellschaftund Romanisationin der riimischen Pruvinz Monia Superior (Budapest-
Amsterdam,1970).
A. MOCSY, Pannonia.
A. MOCSY, 'Pannonia-Forschung 1961-1963',Eirene iv, 1965, 133 ff.
A. MOCSY, 'Pannonia-Forschung 1964-1968',Acta Arch. xxi, 1969, 340 ff.
A. MOCSY, 'Pannonia-Forschung 1969-1972',Acta Arch. xxv, 1973, 375 ff.
T. NAGY, Bp. Miieml.
E. NOWOTNY, 'R6mischeForschungin Osterreich',BRGKxv, 1925, 121 ff.; xvi, 1926, 164 ff.
407
SelectBibliography
P. OLlV A, Pannoniaand the Onsetof Crisis in the RomanEmpire (Prague,1962).
M. PAVAN, 'La Provincia Romanadi PannoniaSuperior', Atli dell' AccademiaNazionale dei
Lincei ccclii, 1955; Memorie viii, 6, 5 (Rome).
B. SARlA, 'RomischeForschungin Jugoslawien',BRGK xvi, 1926, 90 if.
B. SARIA, 'Noricum und Pannonien',Historia i, 195°,436if.
W. SCHMID, 'RomischeForschungin Osterreich',BRGK xv, 1925, 178 if.
A. SCHOBER, Die Riimerzeitin Osterreich(2nd ed., Vienna, 1955).
E. SWOBODA, Carnuntum.
V. VELKOV, Gradat v Trakija i Dakija prez kasnataanticnost(Sofia, 1959).
H. VETTERS, 'Dacia Ripensis', Schriften der Balkankommission,Antiquarische Abteilung xi, I
(Vienna, 195 I).
T. D. ZLATKOVSKAYA, Mezija v i-ii. vekahnaJej ery (Moscow, 1951).
INSCRIPTIONS
A. BA RB, 'Die romischenInschriftendes Burgenlandes',MitteilungendesBurgenlandischenHeimat-

und Naturschutzvereins v, 193I; BurgenlandischeHeimatbliitter i, 1932 if.
L. BARKOCZI, 'Brigetio', Diss. Pann. ii, 22 (Budapest,1944-51).
V. BESEVLIEV, Epigrafskiprinosi (Sofia, 1952).
A. BETZ, 'Die griechischenInschriften Osterreichs',Wiener Studienlxxix, 1966, 597 if.
J. CESKA and R. HOSEK, Inscriptiones Pannoniae Superioris in Slovacia Transdanubianaasservatae
(Brno, 1967).
A. DOBO, 'Inscriptionesad res Pannonicaset Dacicas pertinentesextra fines earundempro-
vinciarum repertae',2nd ed., Diss. Pann. i, I (Budapest,1940)' third ed. forthcoming.
s. DUSANIC, 'Bassianaeand its Territory', Arch. Iugosl. viii, 1967, 67 if.
J. FITZ, 'RomischeInschriften im Komitat Fejer', Alba Regia viii-ix, 1968, 197 if.
B. GEROV, Romanizmat,seeGENERAL WORKS.
v. HOFFILLER and B. SARIA, AI].
Intercisa
A. NEUMANN, 'Inschriften aus Vindobona,'jahrbuch des Vereinsfur Geschichteder Stadt Wien
xvii-xviii, 1962,7if.; xix-xx, 1964, 7 if.
A. VON PREMERSTEIN,N. VULIC and others, jOAIB iii, 1900-XV, 1912.
B. SARIA, 'Die romischenInschriften des Burgenlandes',Burgenl. Heimatb. xiii, 195 I, I if.
A. and J. SASEL, ILjug.
s. SOPRONI, 'RomischeMeilensteineaus Szazhalombatta',Fol. Arch. xxi, 197°,90if.
J. SZILAGYI, Inscr. teg.
E. B. v AGO, 'Neue Inschriften aus Intercisa und Umgebung',Alba Regia xi, 1971, 120 if.
E. VORBECK, Militiirinschriften aus Carnuntum(Vienna, 1954).
N. VULIC, 'Anticki spomenicinasezemlje', Spomeniklxx, lxxv, lxxvii, cxviii, 1931-48.
E. WEBER, 'Die romischenMeilensteineaus dem osterreichischenPannonien',JOAI xlix,
1968-71, 121 if.
Forthcoming: Die RiimischenInschriften Ungarns i (Budapest,1972), ii (1974) etc.
Seealso SCULPTURE IN STONE.
407
SelectBibliography
HISTORY: PRE-ROMAN PERIOD
A. ALFOLDI, 'Zur Geschichtedes Karpatenbeckensim i. Jahrhundertv. Chr.', Archivum

Europae Centro-Orientalis viii, OstmitteleuropiiischeBibliothek xxxvii, 1942.
G. ALFOLDY, 'Des territoires occupespar les Scordisques',Acta Ant. xii, 1964, 107 ff.
G. ALFOLDY, 'Taurisci und Norici', Historia xv, 1966, 224 ff.
B. BENADIK, 'Obraz doby latenskej na Slovensku',Slov. Arch. xix, 1971,465ff.
E. B. BONIS, Die spatkeltischeSiedlungGel!erthegy-Tabdnin Budapest(Budapest,1968).
E. B. BONIS, 'Die Siedlungsverhaltnisseder pannonischenUrbevolkerung-einigeFragen
ihres Weiterlebens',Acta Arch. xxiii, 1971, 33 ff.
J. FILIP, Keltove ve stfedniEvrope(Prague,1956).
o. H. FREY, 'Zur latenezeitlichenBesiedlungUnterkrains',FestschriftDehn 1969, 7 ff.
M. GARASANIN, 'K problematici kasnoglatenau donjem Podunavlji', Naucni Zbornik Matice
Srpske,Odelenjedrufvenih naukexviii, 1958, 92 ff.
M. GARASANIN, 'Contributional'archeologieet al'histoire des Scordisques',A Pedro Bosch-
Gimpera (Mexico, 1963).
M. GARASANIN, 'Ka konfrontacija pisanih i arheoloskihizvora 0 keltima u nasoj zemlji',
Materijali, iii, Simpozijum praistorijske i sredr!fevekovnesekcije arheolofkog druftva ]ugoslavije
Novi Sad196} (Belgrade,1966), I7 ff.
B. GA VELA, '0 etnickim problemimalatenskekulture Zidovara', Starinar xx, 1970, 119 ff.
1. HUNYADY, 'Die Kelten im Karpatenbecken', Diss. Pann. ii, 18 (Budapest,1942-4).
'KeltischeOppida, Symposium1970', Arch. RoZhl. xxiii, 1971, 261-4, with contributionsfrom
B. BENADIK, E. B. BONIS, J. FILIP, J. REITINGER, J. TODOROVIC, v. ZIRRA and others.
F. LOCHNER-HUTTENBACH, 'Illyrier und Illyrisch', Das Altertum xvi, 1970, 216 ff.
M. MACREA, 'Bur~bista §i Celti de la Dunareade mijloc', SCIV vii, 1956, II9 ff.; cpo Dacia
ii, 1958, 143 ff.
z. MARIC, 'Problemedes limites septentrionalesdu territoire illyrien', Symposium1964, 177 ff.
A. MOCSY, 'Die VorgeschichteObermosiensim hellenistisch-romischen Zeitalter', Acta Ant.
xiv, 1966, 87 ff.
A. MOCSY, 'Der vertuschteDakerkriegdes M. Licinius Crassus',Historia xv, 1966, 511 ff.
F. PAPAZOGLU, Sredr!fobalkanska plemenau predrimskodoba (Sarajevo,1969).
P. PETRU, 'VzhodnoalpskiTaurisci', Arh. Vestn. xix, 1968, 339 ff.
P. PETRU, 'Die Hausurnender Latobiker', Situla xi (Ljubljana, 1971).
M. SZABO, The Celtic Heritage in Hungary (Budapest,1971).
J. GY. SZILAGYI, 'Trouvailles grecquessur Ie territoire de la Hongrie', Actesdu viii e Congres
Internationald'ArcheologieClassiqueParis 196} (Paris, 1965), 386 ff.
J. TEJRAL, 'Zur Frage der Stellung Mahrens urn die Zeitrechnungswende', Pam. Arch. lix,
1968, 488 ff.
A. TOCIK, 'K otazceosidlenia juhozapadehoSlovenskana zlome letopoctu', Arch. RoZhl. xi,
1959,841ff.
J. TODOROVIC, Kelti ujugoistocnomEvropi (Belgrade, 1968).
H. VETTERS, 'Zur altestenGeschichteder Ostalpenlander', ]OAI xlvi, 1961-3, 201 ff.
ZS. VISY, 'Die Daker im Gebiet von Ungarn', MFME 1970, 5 ff.
v. ZIRRA, 'Beitragezur Kenntnis des keltischenLa Tene in Rumanien',Dacia xv, 1971, 171 ff.
407
SelectBibliography
Pre-Romancoinage
K. CASTELIN, Germaniaxxxviii, 1960, 32 ff.
E. GOHL, NumismatischeZeitschrift xxxv, 1903, 145 ff.
E. GOHL, Num. KtiZI. i, 19°2, 17 ff.; vi, 19°7,47ff.
A. KERENYI, Fol. Arch. xi, 1959,47ff.
E. KOLNI KOV A, Slov. Arch. xii, 1964, 402 ff.
A. MOCSY, Num. Kijzl.lx-Ixi, 1961-2, 25 ff.
V. ONDROUCH, Keltski mince !ypu Biatec (Bratislava, 1959).
K. PINK, 'Die Miinzpragungder Ostkeltenund ihrer Nachbarn',Diss. Pann. ii, 15 (Budapest,
1939)'
HISTORY: ROMAN PERIOD
A. ALFOLDI, 'The Central DanubianProvinces',CAH xi, 1936.

A. ALFOLDI, 'EpigraphicaIV', Arch. Ert. 1941, 40 ff.
A. ALFOLDI, 'Uber die Juthungeneinfalleunter Aurelian, Serta Kazarovianai', IBAI, xvi,
1950, 21 ff.
A. ALFOLDI, 'The Moral Barrier on Rhine and Danube', The Congress of Roman Frontier
StudiesI949 (Durham, 1952), I ff.
A. ALFOLDI, 'Wo lag das RegnumVannianum?'Siidostforschungen xv, 1956, 48 ff.
G. ALFOLDY, 'Serapis-oltarNyergesujfalub61',Arch. Ert. lxxxviii, 1961, 26 ff.
G. ALFOLDY, 'Der Friedensschluss des Kaisers Commodusmit den Germanen',Historia xx,
1971, 84 ff.
L. BALLA, 'Guerre iazygue aux frontieres de la Dacie', Acta Classica Universitatis de Ludovico
KossuthnominataeDebreceniensisv, 1969, I I 1 ff.
L. BARKOCZI, 'Dak tolmacs Brigetioban',Arch. Ert. 1944-5, 178 ff.
L. BARKOCZI, 'Die Naristenzur Zeit der Markomannenkriege', Fol. Arch. ix, 1957,91ff.
L. BARKOCZI, Intercisa ii, 497 ff.
G. BARTA, 'Lucius Verus and the MarcomannicWars', Acta Classica Universitatis de Ludovico
Kosslflh nominataeDebreceniensisvii, 1971, 67 ff.
A. R. BIRLEY, 'The invasion of Italy in the Reign of Marcus Aurelius', Provincialia, Festschrift
Latfr-Belart (Basel, 1966), 214 ff.
C. DAICOVICIU, Dacica (Cluj, 1970)'
DOBIAS, DCU.
J. DOBIAS, 'Rom und die Volker jenseits der mittleren Donau', Corolla E. Swoboda dicata
(Graz, 1966), 1 I 5 ff.
R. EGGER, 'Ein zweimal beschriebener Weihestein',RAFCi, 312 ff.
J. FITZ, 'Der Besuchdes Septimius Severusin Pannonienim Jahre202', Acta Arch. xi, 1959,
237 ff.
J. FITZ, 'II soggiorno di Caracallain Pannonianel 214', Accademiad'Ungheria, Quaderni di
Documentazioneii, 2 (Rome, 1961).
J. FITZ, 'Die Vereinigung der Donauprovinzenin der Mitte des ;. Jahrhunderts',Stud. Mil.
I 13 ff.
4 10
SelectBibliography
J. F1TZ, 'Pannonienund die Klientelstaatenan der Donau', Alba Regia iv-v, 1965,73if.
J. F1TZ, 'Der markomannisch-quadische Angriif gegen Aquileia und Opitergium', Historia
xv, 1966, 336 if.
J. FITZ, Ingenutlset Rigalien (Brussels,1966).
J. FITZ, 'Zur Geschichteder PraetenturaItaliae et Alpium im Laufe der Markomannenkriege',
Arh. Vestn. xix, 1968, 43 if.
R. GOBL, 'Rex Quadis datus', RheinischesMuseumciv, I96I, 70 if.
F. HAMPL, 'Kaiser Marc Aurel und die Volker jenseitsder Donaugrenze',FestschriftR. Heu-
berger (Innsbruck, 1960), 33 if.
J. HARMATTA, Studiesin the History and Languageof the Sarmatians(Szeged,I972).
H. J. KE L LNE R, 'Raetienund die Markomannenkriege', BcryerischeVorgeschichtsbliitterxxx, I 965,
173 if.
J. KLOSE, RomsKlientelrandstaatenam Rhein und an der Donau (Breslau, I934).
J. K. KOLOSOVSKAYA, 'ZavoevanijaPannoniiRimom', Vestnik drevnei istorii, I96I, I, 60 if.
E. KOSTERMANN, 'Der pannonisch-dalmatische Krieg 6-9 n. Chr.', Hermeslxxxi, I953, 345 if.
A. MOCSY, 'Die Expansionsfrageim i. und ii. Jahrh.', AnnalesUniversitatis Scientiarumde R.
EiHviJs nominatae,Sectio Historica v, I963, 3 if.
A. MOCSY, 'TampiusFlavianusPannoniiban',Arch. Ert. xciii, I966, 203 if.
A. MOCSY, 'Das Geriicht von neuenDonauprovinzenunter Marcus Aurelius', Acta Classica
Universitatis de Ludovico KossuthnominataeDebreceniensisvii, 1971, 63 if.
R. NOLL, 'Zwei unscheinbareKleinfunde aus Emona', Arh. Vestn. xix, I968, 79 if.
c. PATSCH, Beitrage.
H. G. PFLAUM, 'Deux carrieresequestresde Lambeseet de Zana', Lilryca iii, I955, I35 if.
A. VON PREMERSTEIN,'Die Anfange der Provinz Mosien', JOAIB i, 1898, 146 if.
A. VON PREMERSTE1N,'Der Daker- und GermanensiegerM. Vinicius', JOAI xxviii, I933,
140 if.; xxix, 1934, 60 if.
E. RITTERL1NG, 'Die Osi in einer afrikanischenInschrift', Germaniai, 19I7, I32 if.
J. SASEL, 'Bellum Serdicense',Situla iv (Ljubljana, I962).
J. SASEL, 'DrususTi. f. in Emona',Historia xix, I970, 122 if.
J. SASEL, 'Ober Umfang und Dauer der Militarzone PraetenturaItaliae et Alpium zur Zeit
Marc Aurels', Acta of the fth International Congressof Greek and Latin Epigraplij, Cambridge
I9 67 (Oxford, I97I ), 317.
E. SWOBODA, 'Der pannonischeLimes und sein Vorland', Carn.-Jb. I959 (I96I), 17 if.
E. SWOBODA, 'Traian und der pannonischeLimes', Carn.-Jb. I963-4 (I965), 9 if.
R. SYME, 'Lentulus and the Origin of Moesia',JRSxxiv, I934, I33 if.
I. WEILER, 'Huic Severo Pannoniaeet Italiae urbes et Africae contigerunt', Historia xiii,
I9 64, 373 if.
I. WEILER, 'Orbis Romanusund Barbaricum',Carn.-Jb. I963-4 (1965), 94 if.
HISTORY: THE END OF THE ROMAN PERIOD
A. ALFOLDI, Der Untergangder RO'merherrschaJtin Pannonieni-ii (Berlin and Leipzig, I924-6).

L. BARKOCZ1, 'A 6th Century Cemeteryfrom Keszthely-Fenekpuszta', Acta Arch. xx, I968,
275 if.
407
SelectBibliography
I. BONA, 'Ein VierteljahrhundertVolkerwanderungszeitforschung in Ungarn', Acta Arch.
xxiii, 1971, 265 if.
R. EGGER, 'Die ZerstorungPettausdurch die Goten', RAFei, 36 if.
R. EGGER, 'Civitas Noricum', RAFei, II6 if.
R. EGGER, 'Von den letzten RomanenVindobonas',RAFeii, 226 ff.
R. EGGER, 'Historisch-epigraphische Studienin Venetien', RAFei, 45 ff.
A. KISS, 'Pannonialakossaganepvandorlaskorihelybenmaradasanak kerdesehez',Janus Pan-
nonius MuzeumEvkiitryve, 1964 (1965), 81 ff.
H. LADENBAUER-OREL, 'ArchaologischeStadtkernforschung in Wien', Jahrbuch des Vereins
fur Geschichteder Stadt Wien xxi-xxii, 1965-6, 18 ff.
A. MOCSY (review of VARADY, LJP), Acta Arch. xxiii, 1971, 347 ff.
T. NAGY, 'ReoccupationofPannoniafrom the Huns in 427', Acta Ant. xv, 1967, 159 ff.
T. NAGY (review of VARADY, LJP), Acta Ant. xix, 1971, 299 ff.
G Y. SZE KELY, 'Le sort desagglomerations pannoniennesau debutdu moyenageet les origines
de l'urbanisme en Hongrie', Annales Universitatis Scientiarum Budapestinensisde Rolando
Eiitviis nominatae,Sectio His/orica iii, 1961, 59 ff.
L. VARADY, LJP.
H. VETTERS, 'Zur Spatzeit des Lagers Carnuntum', O-rterreichische Zeitschrift fur Kunst tllld
Denkmalpflegexvii, 1963, 15 7 if.
ADMINISTRATION
L. BALLA, 'Die Inschrift eines Senatorsaus Savaria', EpigraphischeStudieniv (Cologne-Graz,

1967), 61 f.
A. BETZ, 'Zum Sicherheitsdienst in den Provinzen',JOAIB xxxv, 1943, 137 ff.
A. R. BIRLEY, Acta Antiqua Philippopolitana 1963, 107 if.
A. DOBO, 'Publicum portorium Illyrici', Diss. Pann. ii, 16 (Budapest,1940).
A. DOBO, Verwaltung.
R. EGGER, 'Das Praetorium als Amtssitz und Quartier romischer Spitzenfunktionare',
Sitzungsberichteder OsterreichischenAkademie,Philosophisch-historischeKlasse ccl/4 (Vienna,
1966).
J. FITZ, 'The Governorsof PannoniaInferior', Alba Regia xi, 1971, 145 ff.
J. FITZ, 'Le iscrizioni del Capitolium di Gorsium', Rivista Storica dell'Antichitd I, 1971, 145 if.
J. FITZ, 'A pannoniaibanyakigazgatasa',Alba Regia xi, 1971, 154 ff.
T. NAGY, 'Zu den Militar- und VerwaltungsreformenDiocletians im pannonischenRaum',
Akte des Iv. Internationalen Kongressesfur griechische und lateinische Epigraphik, Wien, I962
(Vienna, 1964), 274 ff.
A. STEIN, 'Die Legatenvon Moesien', Diss. Pann. i, II (Budapest,1940).
R. SYME, 'GovernorsofPannoniaInferior', Historia xiv, 1965, 342 ff.
R. SYME, 'Hadrian and Moesia', Arh. Vestn. xix, 1968, 101 ff.
I. WEILER, 'Beitrage zur Verwaltung Pannonienszur Zeit der Tetrarchie',Situla viii (Ljub-
ljana, 1965), 141 ff.
412
SelectBibliography
MUNICIPALITIES
G. ALF OLDY, 'Augustalen- und Sevirkorperschaftenin Pannonien',Acta Ant. vi, 1958,

433 ff.
G. ALFOLDY, 'Eine Strassenbauinschrift aus Salona',Acta Arch. xvi, 1964, 247 ff.
G. ALFOLDY, 'Municipium Iasorum', Epigraphica xxvi, 1965, 95 ff.
s. DUSANIC, 'Bassianaeand its Territory', Arch. lugosl. viii, 1967,67ff.
J. FITZ, 'Angaben zu den Gebietsveranderungen der Civitas Eraviscorum', Acta Arch.
xxiii, 1971, 47 ff.
M. MIRKOVIC, RGD.
A. MOCSY, 'Zur Geschichteder peregrinenGemeindenin Pannonien',Historia vi, 1957,488ff
A. MOCSY, 'Scribak a pannoniaikisvarosokban',Arch. Ert. xci, 1964, 16.
A. MOCSY, 'Decurio Eraviscus',Fol. Arch. xxi, 1970, 59 ff.
v. VELKOV, 'Ratiaria', Eirene v, 1966,155ff.
MILITARY TERRITORIES
L. BARKOCZI, 'Beitrage zum Rang der Lagerstadtam Ende des ii. und am Anfang des iii
Jhs.', Acta Arch. iii, 1953, 201 ff.
R. EGGER, 'Bemerkungenzum Territorium pannonischerFestungen',RAFCii, 135 ff.
A. MOCSY, 'Das Territorium Legionis und die Canabaein Pannonien' Acta Arch. iii, 1953,
179 ff.
A. Moe SY,'II problemadelle condizioni del suolo attribuito aIle unita militari nelle province
danubiane',Atti del Convegnolnternazionalesui diritti locali nelle province dell'lmpero Romano,
Roma, I9P (Rome, forthcoming).
F. VITTINGHOFF, 'Die Bedeutung der Legionslager flir die Entstehung der romischen
Stadtean der Donauund in Dakien', Studienzur europiiischenVor- und Friihgeschichte(Neu-
miinster, 1968), 132 ff.
THE ARMY
G. ALFOLDY, 'Die Truppenverteilungder Donaulegionenam Endedesi. Jhs.',Acta Arch. xi,

1959, 122 ff.
G. ALFOLDY, 'Thrakischeund illyrische Soldatenin den rheinischenLegionen',Epigraphische
Studieniv, 1967, 26 ff.
L. BARKOCZI, 'A new military diploma from Brigetio', Acta Arch. ix, 1959,413ff.
A. BETZ, 'Zur Geschichteder Legio x. Gemina', Corolla E. Swobodadedicata (Graz, 1966),
39 ff.
s. DUSANIC, 'Rimskavojska na istocnom Sremu',ZFF x, 1968, 107 ff.
J. FITZ, 'Legati legionum PannoniaeSuperioris',Acta Ant. xi, 1961, 159 ff.
J. FITZ, 'A Military History ofPannoniafrom the MarcomannicWars to the Deathof Alexan-
der Severus',Acta Arch. xiv, 1962, 25 ff.
413
SelectBibliography
J. FITZ, 'Massnahmenzur militarischen Sicherheitvon PannoniaInferior unter Commodus',
Klio xxxix, 1961, 199 ff.
J. FITZ, 'Reorganisationmilitaire au debut des guerresmarcomannes',Hommagesd M. Renard
ii (Brussels,1969), 262 ff.
B. GEROV, 'EpigraphischeBeitrage zur Geschichtedes mosischenLimes in vorclaudischer
Zeit', Acta Ant. xv, 1967, 91 ff.
o. V. KUDRIAVCEV, 'Dunaiskielegioni i ih znaceniev istorii rimskoi imperii', Issledovanijapo
istorii balkansko-dunaiskihoblasti (Moscow, 1957), 147 ff.
M. MIRKOVI<:' 'Cohorsi. Cantabrorumi posadakastelaAquae', ZFF viii/I, 1964,87fr.
M. MIRKOVIC, 'Die Auxiliareinheiten in Mosien unter den Flaviern', EpigraphischeStudien
v, 1968, 177 ff.
A. MOCSY, 'Zur frlihesten Besatzungsperiode in Pannonien',Acta Arch. xxiii, 1971,41fr.
T. NAGY, 'The Military Diploma of Albertfalva', Acta Arch. vii, 1956, 17 fr.
T. NAGY, 'Commandersof Legions in the Age of Gallienus', Acta Arch. xvii, 1965, 295 fr.
H. NESSELHAUF, 'Zwei Inschriften aus Belgrad', ZA x, 1960, 191 fr.
M. PAVAN, 'Iscrizioni latine ad Albano Laziale', Athenaeumxl, 1962, 85 ff.
D. PROTASE, 'La legion IV Flavia au nord du Danubeet la premiereorganisationde la Dacie
romaine',Acta of the fth International Congressof Greekand Latin Epigrapllj, Cambridge 1967
(Oxford, 1971), 337 fr.; cpo I. GLODARIU, ibid., 327 fr.
A. RADNOTI and L. BARKOCZI, 'The Distribution of Troopsin PannoniaInferior during the
2nd Century A.D.', Acta Arch. i, 195 I, 191 fr.
A. RADNOTI, 'Zur Dislokation der Auxiliartruppen in den Donauprovinzen',Limes-Studien
(Basel, 1957), 142 ff.
S. SOPRONI, 'Two Inscribed Relics of the Cohors XlIX Vol. c. R.', Fol. Arch. xvi, 1964,
33 fr.
s. SOPRONI, 'Der Stempelder Legio XIV Geminain Brigetio', Fol. Arch. xvii, 1965, II9 fr.
R. SYME, 'The First Garrisonof Trajan'sDacia', Laur. Aqu. i, 267 fr.
E. TOTH and G. VEKONY, 'Beitragezu PannoniensGeschichteim Zeitalter des Vespasianus',
Acta Arch. xxii, 1970, 133 ff.
D. VUCKOVIC-TODOROVIC, 'Vojnicka diploma iz kastrumaTaliatae', Starinar iV/18, 1967,
21 ff.
K. WAC HTEL, 'Kritisches und Erganzendeszu neuen Inschriften aus Mainz', Historia xv,
1966, 247 fr.
ETHNIC COMPOSITION, SOCIAL AND ECONOMIC HISTORY
M. R. ALFOLDl, 'Der Geldverkehrin Intercisa',Intercisa i, 142 fr.

G. ALFOLDY, 'Municipalis kozepbirtokAquincum kornyeken',Ant. Tan. vi, 1959, 19 fr.
G. ALFOLDY, 'Die Valerii in Poetovio', Arh. Vestn. xv-xvi, 1964-5, 137 fr.
L. BALLA, 'Ostliche ethnischeElementein Savaria',Acta Arch. xv, 1963, 225 fr.
L. BALLA, 'Gesellschaftund Geschichtevon Savaria',Die riJ"mischen Steindenkmiilervon Savaria
(Budapest,1971), 19 fr.
L. BALLA and I. TOTH, 'A propos des rapportsentre la Pannonieet la Dacie', Acta Classica
Universitatis Scientiarumde Ludovico KossuthnominataeDebreceniensisiv, 1968, 69 fr.
414
SelectBibliography
L. BARKOCZI, 'The Populationof Pannoniafrom Marcus Aurelius to Diocletian', Acta Arch.
xvi, 1964, 257 if.
O. DAVIES, RomanMines in Europe (Oxford, 1935).
s. DUSANIC, 'Novi antinojev natpisi Metalla Municipii Dardanorum',ZA xxi, 1971, 291 if.
G. ELMER, 'Der romische Geldverkehrin Carnuntum', NumismatischeZeitschrift lxvi, 1933,
55 if.
J. FITZ, 'Die domus Heraclitianain Intercisa', Klio I, 1968, 159 if.
J. FITZ, Les.ryriens d lntercisa (Brussels,1972).
s. FOLTINY, 'Eine dakischeHenkelschaleaus Mullendorf', Barb-Festschr.79 if.
F. FREMERSDORF,'RheinischerExport nach dem Donauraum',Laur. Aqu. i, 168 if.
J. FROMOLS, 'Decouverted'une plaque danubiennea Port sur Saone',Jahrbuch des Romisch-
GermanischenZentralmuseums v, 1958, 259 if.
D. GABLER, 'Munera Pannonica',Arch. Brt. xciii, 1966, 20 if.
J. GARBSCH, Die norisch-pannonischeFrauentracht(Munchen, 1965).
E. GREN, Der Miinzfund von Viminacium (Uppsala-Leipzig,1934).
E. GREN, Kleinasien und der Ostbalkanin der wirtschaftlichen Entwicklungder rijmischen Kaiserzeit
(Uppsala, 1941).
R. KA TIeIC, 'Die neuesten Forschungenuberdie einheimischeSprachschichtin denillyrischen
Provinzen',SymposiumI964, 31 if.
R. KATIeIC, 'Zur Frage der keltischen und pannonischenNamengebieteim romischen
Dalmatien',Godifnjak iii, Centarza BalkanoloskaIspitivanja i (Sarajevo,1965), 53 if.
R. KA TIeIC, 'Keltska osobnaimena u anticko Sloveniji', Arh. Vestn. xvii, 1966, 145 if.
H. G. KOLBE, 'Die Laufbahn des Faustinianusaus Carnuntum', Carn.-Jb. 1963-4 (1965),
48 if.
J. K. KOLOSOVSKAYA, 'Veteranskoezemlevladeniev Pannonii', Vestnik drevnei istorii 1963,
96 if.
10 MIKL-CURK, 'Gospodarstvona ozemlju danasnjeSlovenije v zgodnji antiki', Arh. Vestn.
xix, 1968, 307 if.
A. MOCSY, Beviilk.
A. MOCSY, 'Dekentioi,' Ant. Tan. xiii, 1966,242if.
A. M 0 C SY, 'V orarbeitenzu einem Onomasticonvon Moesia Superior', Godilniak iii, Centar
za BalkanoloskaIspitivanja 6 (Sarajevo,1970), 139 if.
A. MOCSY, 'Die lingua Pannonica',SymposiumI967, 195 if.
GY. NOVAK1, 'Uberrestedes Eisenhuttenwesen in Westungarn',Barb-Festschr.163 if.
v. ONDROUCH, Ndlezykeltskych,antickycha lryzantskychmind na Slovensku(Bratislava, 1964).
S. PANCIERA, La vita economicadi Aquileia (Venice, 1957).
c. PA TS C H, 'Die Saveschiifahrtin der Kaiserzeit',JOAl, viii, 19°5, 139 if.
c. PATSCH, 'Zur Geschichtevon Sirmium', Sirena Buliciana (Split, 1924), 229 if.
T. PEKARY, 'Aquincum penzforgalma',Arch. Brt. lxxx, 1953, 106 if.
K. PINK, 'Der Geldverkehram osterreichischenDonaulimes',Jahrbuch fiir Landeskundevon
Niederiisterreichxxv, 1932,49if.
M. I. ROSTOWZEW, 'Ein Speculatorauf der Reise. Ein Geschaftsmann bei der Abrechnung',
RomischeMitteilungen xxvi, 19II, 278 if.
J. SASEL, 'Caesernii',ZA x, 1960, 201 if.
J. SASEL, 'Barbii', Eirene v, 1966, II7 if.
407
SelectBibliography
J. SASEL, 'Keltisches Portorium in den Ostalpen', Corolla E. Swobodadedicata (Graz, 1966),
198 ff.
s. SOPRON1,'tiber den Mlinzumlauf in Pannonienzu Ende des 4. Jhs.', Fol. Arch. xx, 1969,
69 ff.
E. M. STAERMAN, 'Etniceski i socialni sostav rimskogo voiska na Dunae', Vestnik drevnei
istorii 1946, 3, 256 ff.
M. SZABO, 'Nehanynyelveszetiszemponta pannoniaikelta szemelynevanyagvizsgalataban',
Ant. Tan. x, 1963, 220 ff.
M. SZABO, 'A pannoniaikelta szemelynevanyag vizsgalata',Arch. Ert. xci, 1964, 165 ff.
J. SZILAGYI, 'Belgaeim Vorort von PannoniaInferior', Hommagesd M. Renardii (Brussels,
1969), 708 ff.
1. TOTH, 'Eine Tempelbauinschrifteines ritterlichen Dekurios aus Brigetio', Acta Classica
Universitatis Scientiarumde Ludovico KossuthnominataeDebreceniensisvii, 1971, 91 ff.
1. TOTH, 'A savariaiIuppiter Dolichenus-szentely feliratos emlekeirol',Arch. Ert. xcviii, 1971,
80 ff.
GY. UROGDI, 'A bankeletnyomai Aquincumban',Bp. Reg. xxi, 1964,239ff.
v. VELKOV, 'Kleinasiatenund Syrier in den Balkangebieten',EtudesHistoriquesde XIIe Congres
International des SciencesHistoriquesd Vienne I96J (Sofia, 1965), 25 ff.
P. VEYNE, 'Epigraphica',Latomusxxiii, 1964, 30 ff.
J. WIELOWIEJSK1, Kontakty Noricum i Pannonii Z ludami prJlnoarymi (Warsaw, 1970).
MINTS
Viminacium
S. DUSANIC, Starinar iV/12, 1961, 141 ff.
J. FITZ, Num. KiiZI. lxii-lxiii, 1963-4, 19 ff.
Siscia
A. ALFOLDI, 'Vorarbeitenzu einem Corpus der in Siscia gepragtenRomermlinzen',Num.
KiiZI. xxvi, 1927-xxxix, 1940.
V. LANYI, 'The Coinageof Valentinian I in Siscia', Acta Arch. xxi, 1969, 33 ff.
RELIGION
Paganism
G. A LFOLDY, 'Geschichtedes religiosen Lebensin Aquincum', Acta Arch. xiii, 1961, 103 ff.
G. ALFOLDY, 'Ein Denkmal des Isis-Sarapis-Kultesin Pannonien',Alba Regia iv-v, 1965,
87 ff.
L. BALLA, 'Deorum Prosperitati', Acta Classica Universitatis Scientiarum de Ludovico Kossuth
nominataeDebreceniensisii, 1966, 96 ff.
407
SelectBibliography
L. BALLA, 'Religion und religioses Leben', Die romischenSteindenkmiilervon Savaria(Budapest,
1971),37ff.
ZS. BANKI, 'A tad Iuppiter-SilvanusDomesticus-oltar',Alba Regia vi-vii, 1967, 165 ff.
A. CERMANOVIC-KUZMANOVIC, 'Die Denkmaler des thrakischenHeros in Iugoslavien',
Arch. lugosl. iv, 1963, 3I ff.
A. CERMANOVIC-KUZMANOVIC, 'Nekoliko spomenikatrackog konjanika',Starinar iV/13-1,
1965, I 14 ff.
A. DOBROVITS, 'Az egyiptomi kultuszok emh~kei Aquincumban',Bp. Reg. xiii, 1943,45 ff.
F. FULEP, 'New Remarkson the Synagogaat Interdsa',Acta Arch. xviii, 1966, 93 ff.
Z. KADAR, Die kleinasiatisch-ryrischenKulte zur Ro"merzeit in Ungarn, Etudes Preliminaires ii
(Leiden, 1962).
G. I. KAZAROW, 'Denkmalerdes Dolichenus-Kultes',JOAl xxvii, 1932, 168 ff.
V. KOLSEK, 'Pregledantickih kultov na slovenskern ozemlju', Arh. Vestn. xix, 1968, 273 ff.
N. LANG, 'Das Dolichenumvon Brigetio', Laur. Aqu. ii, 165 ff.
N. LANG, 'A savariaiDolichenus-csoportozat', Arch. Ert. 1943, 64 ff.
N. LANG, 'Die Dolichenus-Votivhanddes BudapesterNationalmuseums',Arch. Ert. 1946-8,
183 ff.
R. MARIe, Anticki kultovi u nasqj zemiji (Belgrade, 1933).
T. NAGY, 'Vallasi elet Aquincumban',Bp. Tod. 386 ff.
T. NAGY, 'A sarkesziMithraeum', Bp. Reg. xv, 195°,46ff.
T. NAGY, 'Das Mithras-Reliefvon Paks',Acta Ant. vi, 1958, 407 ff.
D. PINTEROVIC, 'Da li je u rimsko koloniji Mursi postojalosinagoga?',Olj. Zb. ix-x, 1965,72ff.
G. SEURE, 'Votivni reljefi u beogradskommuzeji', Starinar iii/I, 1922-3, 238 ff.
R. M. SWOBODA, 'Denkmaler des Magna-Mater-Kultesin Slowenien und Istrien', Bonner
jahrbiicher clxix, 1969, 195 ff.
T. SZENTLELEKY, 'Das Iseum von Szombathely',Neue Beitriige zur Geschichteder Alten Welt
ii (Berlin, 1965), 381 ff.
T. SZENTLELEKY, Das lsis-Heiligtum von Szombathefy(Szombathely,1965).
M. SZOKE, 'Building Inscription of a Silvanus-Sanctuary from Cirpi', Acta Arch. xxiii, 1971,
221 ff.
I. TOTH, 'MegjegyzesekMithras pannoniaikultuszanaktortenetehez',Ant. Tan. xii, 1965, 86 ff.
V. WESSETZKY, Die agyptischenKulte zur Riimerzeit in Ungarn, Etudes PrBliminaires i (Leiden,
1961).
v. WESSETZKY, 'Der Isis-Altar von Sopron', Das Altertum x, 1964, 154 ff.
K. WIGAND, 'Die Nutrices Augustaevon Poetovio',jOAlB xviii, 1915,189ff.
E. WILL, 'Les fideles de Mithra a Poetovio', Adriatica, MiscellaneaGrga Novak (Zagreb, 1970),
633 ff.
L. ZOTOVIC, 'Kult JupiteraDepulzora',Starinar iV/17, 1966,43ff.
L. ZOTOVIC, Les cultes orientaux sur Ie territoire de la Mesie Superieure,Etudes Preliminaires xi
(Leiden, 1966).
Christianity
A. ALFOLDI, Tracce del cristianesimonell'epocadelle grandi migrazioni in Ungheria (Rome, 1938).

R. EGGER, Der heilige Hermagoras(Klagenfurt, 1948).
407
SelectBibliography
L. NAGY, PannoniaSacra, SzentIstvdn EmJekktinyvi (Budapest,1938).
T. NAGY, 'A pannoniaikeresztenysegtortenete',Diss. Pann. ii, 12 (Budapest,1939).
R. NOLL, PruhesChristentumin Osterreich (Vienna, 1954).
H. VETTERS, 'Drei Silber1o£Ielaus Carnuntum',BRGKxlix, 1968, 149 £I.
J. ZEILLER, Les origines chretiennesdans les provincesdanubiennesde I'Empire Romain(Paris, 1918).
(seealso GRAVES AND CEMETERIES: INHUMATION)
CULTURE, EDUCATION, LATIN
S. DU5ANIC, 'Cuillula, an Epigraphic Hapax Legomenon', Epigraphische Studien v, 1968,

158 £I.
R. EGGER, 'EpikureischeNachklangeaus Aquincum', RAPC ii, 153 £I.
A. HEKLER, 'Kunst und Kultur Pannoniensin ihren Hauptstromungen',Strena Buliciana
(Split, 1924), 107 £I.
J. HERMAN, 'Posit = posuit et questionsconnexesdansles inscriptionspannoniennes',Acta
Ant. ix, 1961, 321 £I.
J. HERMAN, 'Latinitas Pannonica',Pilolrfgiai KijZloi!J 1968, 364 £I.
H. M1HAESCU, Limba latina In provinciile dunareneale imperiului Roman(Bucharest,1960).
L. NAGY, AZ aquincumiorgona (Budapest,1934).
J. GY. SZILAGYI, 'Megjegyzesekaz uj szentendreiversesfeliratr6l', Arch. Brt. xc, 1963, 189 £I.
w. WALC KER-MAYER, Die rijmische Orgel von Aquincum(Stuttgart, 1970).
THE ARCHAEOLOGICAL MATERIAL
Coarsepottery
E. BONIS, 'Die kaiserzeitlicheKeramik von Pannonien',i, Diss. Pann. ii, 20 (Budapest,1942).
E. BONIS, Pol. Arch. xxi, 1970, 71 £I.
P. PETRU, Razprave,SlovenskaAkademijavi (Ljubljana, 1969), 197 £I.
K. sz. POCZY, Intercisa ii, 29 £I.
K. sz. POCZY, Arch. Ert.lxxxii, 1955, 56 £I.
K. sz. POCZY, Acta Arch. vii, 1956, 73 £I.
A. SALAMON, Pol. Arch. xx, 1969, 53 £I.
A. SCHORGENDORFER,Die ro·merzeitlicheKeramik in den Ostalpenliindern(Vienna, 1942).
U. TRINKS, Carn.-Jb. 1957, 51 £I.; 1959, 70 £I.
B. VIKIC-BELANCrC, Starinar iV/13-I4, 1965, 89 £I.
I. WELLNER, Arch. Brt. xcii, 1965, 42 £I.
Lamps
D. GAJ-POPOVIC,ZNMB iii, 1963, 129 £I.
D. IV ANY!, 'Die pannonischenLampen', Diss. Pann. ii, 2 (Budapest,1935).
V. KONDIC and J. TODOROVIC, GodiI'!iak Muzeja Grada Beogradaiii, 1956, 63 £I.
4 18
SelectBibliography
A. NEUMANN, 'Lampen und andereBeleuchtungsgerate
aus Vindobona', RLiO xxii (Vienna,
1967).
T. SZENTLELEKY, Bp. Reg. xix, 1959, 167 ff.
Kilns, potteries
L. BARKOCZI, Pol. Arch. viii, 1956, 81 ff.
B. KUZSINSZKY, Bp. Reg. xi, 1932.
J. MEDUNA, Germaniaxlviii, 1970, 44.
L. NAGY, Arch. Brl. 1942, 162 ff.
GY. PARRAGI, Arch. Erl. xcviii, 1971, 60 ff.
z. SUBI(";, Arh. Vestn.,xix, 1968,455ff.
Sigil/ata (samianware)
F. EICHLER, Germaniaxxv, 1941, 30 ff.
D. GABLER, Arrabona vi, 1964, 5 ff.; ix, 1967, 21 ff.
D. GABLER, Arch. Brt. xci, 1964,94ff.; xcv, 1968, 2II ff.
D. GABLER, Acta Ant. xvi, 1968, 297 ff.
D. GABLER, Acta Arch. xxiii, 1971, 83 ff.
D. GABLER, Btryerische Vorgeschichtsbkitterxxxi, 1966, 123 ff.; xxxiii, 1968, 100 ff.
G. JUHASZ, 'Die Sigillaten von Brigetio', Diss. Pann. ii, 3 (Budapest,1935).
G. JUHASZ, Arch. Brt. xlix, 1936, 33 ff.
F. KRIZEK, Slov. Arch. ix, 1961, 301 ff.; xiv, 1966,97ff.
I. MIKL-CURK, Terra sigil/ata in sorodnevrste keramikeiZ Poetovija (Belgrade-Ljubljana,1969).
L. NAGY, Bp. Reg. xiv, 1945, 303 ff.
B. RUTKOWSKI, WiadomosciArcheologicznexxx, 1964,75 ff.
B. RUTKOWSKI, Archeologiaxviii (Warsaw, 1967), 55 ff.
v. SAKAR, Arch. RoZhl. xxi, 1969, 202 ff.
J. TEJRAL, Arch. RoZhl. xxii, 1970, 389 ff.
B. VIKIC-BELANCIC, Arh. Vestn. xix, 1968, 509 ff.
zs. VISY, Alba Regia x, 1969, 87 ff.
The potter Pacatus

A. ALFOLDI, Pol. Arch. i-ii, 1939, 97 ff.
K. KISS, Laur. Aqu. i, 188 ff.
Terracotta
A. ALFOLDI, Laur. Aqu. i, 312 ff.
L. CASTIGLIONE, Actesdu viii e CongresInternational d'Archeologie Classique,Paris, I96J (Paris,
1965), 361 ff.
A. CERMANOVIC-KUZMANOVIC, ZNMB iv, 1964, 141 ff.
M. VELIC KOVIC, Katalog grlkih i rimskih terakota. Narodni MuZq, Antika iii (Belgrade,1957).
419
SelectBibliography
Bronze
zs. BANKI, 'Objets romains figures en bronze, argent et plomb', Collection du Mush Roi Saint
Etienne (Szekesfehervir,1972).
D. GASPAR, 'Spfltromische Kastchenbeschlflgein Pannonien',Acta Antiqua et Archaeologica
Universitatis de Attila JOzsefnominataeXV!I-2 (Szeged,1971).
1. KOVRIG, 'Die Haupttypender kaiserzeitlichenFibeln in Pannonien',DiS!. Pann. ii, 4 (Buda-
pest, 1937).
A. MOTYKOV A-sNEIDRovA, 'Norisch-pannonische Giirtelbeschlageund ihre Nachbildungen
in Bohmen',Pam. Arch. lv, 1964, 350 if.
E. PATEK, 'Verbreitungund Herkunft der romischenFibeltypenvon Pannonien',DiS!. Pann.
ii, 19 (Budapest,1942).
Lj. POPOVIC, DJ. MANO-ZISI, M. VELICKOVIC and others, Greek, Roman, and Barfy
Christian Bronzesin Yugoslavia(Belgrade, 1969)'
A. R A D N 0 TI, Bronzegefiisse.
A. RADNon, Intercisa ii, 173-364.
I. SELLYE, 'Les bronzesemaillesdelaPannonie •• Diss.Pann.ii.8 (Budapest,1939)'
I. SELLYE, 'A pannoniaiattort femmunkik ittekintese',Arch. Ert. 1940,236if.; 1941,62if.
I. SELLYE, 'Recueil des bronzes ajoures en Pannonie',Hommagesd M. Renard iii (Brussels
1969), 516 if.
1. SELLYE, 'Adatok az arrabonaifemmiivesseghez', Arrabona xii, 1970, 69 if.
Sculpturein bronze
I. BONA, Acta Arch. xxiii, 1971, 225 if.
R. FLEISCHER, Die romischenBronzenaus Osterreich(Mainz, 1967).
G. GAMER, Germaniaxlvi, 1968, 53 if.
Z. KADAR, Fol. Arch. xvii, 1965, I I I if.
1. PAULOVICS, Pannoniai, 1935, 21 if.
D. P1NTEROVIC, Oij. Zb. viii, 1962, 117 if.; ix-x, 1965,92if.
T. SZENTLELEKY, VMMK iv, 1965, 103 if.
Glass
L. BARKOCZ1, Fol. Arch. xviii, 1966,67if.; xix, 1968, 59 if.; xx, 1969,47if.; xxii, 1971, 71 if.
L. BARKOCZI and A. SALAMON, Arch. Ert. xcv, 1968, 29 if.
A. BENKO, 'Uvegcorpus',RegesZetiFiizetek ii, I I (Budapest,1962).
i. BONIS, Bp. Reg. xiv, 1945, 561 if.
F. FULEP, Acta Ant. xvi, 1968, 401 if.
K. M. KABA, Bp. Reg. xviii, 1958,425if.
s. PETRU, Razprave,SlovenskaAkademijavi (Ljubljana, 1969), 163 if.
A. RADNon, Intercisa ii, 141 if.
R. SUNKOWSKY, Antike Gliiser in Carnuntumund Wien (Vienna, 1956).
420
SelectBibliography
Sculpturein stone
L. BALLA, T. P. BUOCZ, Z. KADAR, A. MOCSY and T. SZENTLELEKY, Die riimischen Steindenk-
maier von Savaria (Budapest,1971).
J. BRUNS MID, 'Kameni spomenici Hrvatskoga Narodnoga Muzeja u Zagrebu', VHAD
vii, 1903-xi, 1911.
A. SZ. BURGER, 'Aldozati jelenet Pannoniakoemlekein', RegeszetiFiizetek ii, 5 (Budapest,
1959)·
A. SZ. BURGER, 'Collegiumi kOfaragomiihelyekAquincumban',Bp. Reg. xix, 1959, 9 ff.
A. CERMANOVIC-KUZMANOVIC, 'Die dekoriertenSarkophage in den romischenProvinzen
von Jugoslawien',Arch. Iugosl. vi, 1965, 89 ff.
E. DIEZ, 'Die Aschenkistenvon Poetovio',]OAI xxxvii, 1948, 151 ff.
E. DIEZ, 'Der Giebel des carnuntinischenFahnenheiligtums',Corolla E. Su/oboda dedicata
(Graz, 1966), 105 ff.
D. P. DIMITROV, Nadgrobniteploci ot rimsko vreme v severnaBalgarija (Sofia, 1942).
G. ERDELYI, 'Steindenkmaler', Intercisa i, 169 ff.
G. ERDELYI, 'Adatok a pannoniaisiraediculakhoz',Arch. Ert. lxxxviii, 1961, 184 ff.
D. GABLER, 'Arrabonakornyekenekkoplasztikai emlekei', Arrabona x, 1968, 51 ff.
D. GABLER, 'Scarbantiakornyekenekkoplasztikaiemlekei', Arrabona, xi, 1969, 5 ff.
M. GORENC, 'Klesarnakiparska manufakturau nasim krajima', Arh. Vestn. xix, 1968, 195 ff.
G. KRA STEVA-NOZAROV A, Ukazatelna predmetiizobrazenivarhu anticni pametniciot Balgarija
(Sofia, 1958).
M. L. K RUG E R, 'Die Rundskulpturendes Stadtgebietes von Carnuntum',Corpusder Skulpturen
der riimischen Welt, Osterreichi, 2 (Vienna, 1967).
M. L. K RUG E R, 'Die Reliefs des Stadtgebietes von Carnuntum,i, Die figiirlichen Reliefs',
Corpusder Skulpturender riimischen Welt, Osterreichi, 3 (Vienna, 1970); 'ii, Die Dekorativen
Reliefs', ibid., i. 4 (Vienna, 1972).
A. MOCSY, Fol. Arch. ix, 1957, 83 ff.; xvi, 1964,43ff.
L. NAGY, Laur. Aqu. ii, 232 ff.
T. NAGY, 'KOfaragases szobniszatAquincumban',Bp. Reg. xxii, 1971, 103 ff.
A. NEUMANN, 'Die Skulpturendes Stadtgebietes von Vindobona', Corpus der Skulpturender
riimischen Welt, Osterreichi, I (Vienna, 1970).
D. PINTEROVIC, Olj. Zb. xi, 1967.
A. SCHOBER, Die romischenGrabsteinevon Noricum und Pannonien(Vienna, 1923).
(Seealso INSCRIPTIONS)
Mythological scenes
E. DIEZ, Carn.-]b. 1963-4(1965),43ff.
G. ERDELYI, Arch. Ert.lxxvii, 1950,72ff.
G. ERDELYI, Acta Ant. xiv, 1966, 211 ff.
G. ERDELYI, Acta Arch. xiii, 1961, 89 ff.
Z. KADAR, Actesdu viii e CongresInternational d'Archiologie Classique,Paris IgtJ) (Paris, 1965),
381 ff.
421
SelectBibliography
Wall-paintings, mosaics,stuccos
K. M. KABA, Bp. Reg. xvi, 1955, 255 ff.
H. KENNER, La Mosai'quegreco-romain(Paris, 1965), 89 ff.
A. K ISS, ibid., 297 ff.
DJ. MANO-ZISI, ibid., 290 ff.
L. NAGY, Arch. Brt. xli, 1927, II4 ff.
L. NAGY, RijmischeMitteilungen xl, 1925, 51 ff.; xli, 1926, 79 ff.
1. WELLNER, Acta Arch. xxi, 1969, 235 ff.
I. WELLNER, 'A magyarorszagiromaikori epuletekbelsa diszita muveszete',Bpites es Bpite-
szettudomdtryii, 1971, 327 ff.
Art in general: variousfinds and treasures

E. BONIS, aus Brigetio', Fol. Arch. xix, 1968, 25 ff.
'Emaillierte Palastragerate
S. FERRI, ARD.
M. GRBIC, Plastika.
Z. KADAR, 'Adatok a Duna videki kesoantikezustedenyek problematikajahoz',Fol. Arch. xii,
1960, 133 ff.
M. MANOJLOVIC-MARIJANSKI, Kasnorimskil!femovi iZ Berkasova(Novi Sad, 1964).
D J. MANO-ZISI, Antika u NarodnomMuzeju u Beogradu(Belgrade,1956).
DJ. MANO-ZISI, Nalaz iZ Tekije (Belgrade, 1957)'
R. NOLL, Kunst der Rijmerzeitin Osterreich (Salzburg,1949)'
v. ONDROUCH, Bobdtehrolry Z dolry rimskej-na Slovensku(Bratislava, 1957).
D. PINTEROVIC, 'Gemes terenaMurse', Og. Zb. ix-x, 1965, 52 ff.
K. SZ. POCZY, 'GemaltesMannerportrataus einem Mumiengrabin Aquincum', Studien zur
Geschichteund Philosophiedes Altertums(editor J. HARMATTA) (Budapest,1968), 331 ff.
K. B. SEY, M. KAROLY I and T. SZENTLELEKY, 'A balozsameggyesi romai ekszer-es eremlelet',
Arch. Brt. xcviii, 1971, 190 ff.
E. B. THOMAS, 'SeverischesGoldgewebeaus Viminacium', Studien zur Geschichteund Philo-
sophiedes Altertums(editor J. HARMATT A) (Budapest,1968), 337 ff.
E. B. THOMAS, Helme, Schilde, Dolche (Budapest,1971).
Gravesand cemeteries
Cremation
E. BONIS, Fol. Arch. xii, 1960, 91 ff.
A. SZ. BURGER, Arch. Brt. xciii, 1966, 254 ff.
D. DZONOVA, Arheologija iv (Sofia, 1962), 3, 32 ff.
F. FULEP, Acta Arch. ix, 1958,373ff.
M. V. GARA SAN IN, GodiSnjak,vi, AkademijaNauka i UmjetnostiBosnei Hercegovine,Centarza
BalkanoloskaIspitivanja, 4 (Sarejevo,1968), 5 ff.
T. KNEZ, Arh. Vestn. xix, 1968, 221 ff.
422
SelectBibliography
T. KNEZ, P. PETRU and s. PETRU, Razprave,SlovenskaAkademijavi (Ljubljana, 1969), 7 ff.,
85 ff., 109 ff.
A. MOCSY, Arch. Ert. lxxxi, 1954, 167 ff.
E. F. PETRES,Pol. Arch. xvii, 1965, 96 ff.
A. RADNOTI, Barb-Pestschr.199 ff.
A. SCHOBER,](jAIB xvii, 1914, 222 if.
D. SREJOVI<::, Starinar iV/13-I4, 1965, '"52 ff.
1. WEILER, Carn.-]b. 1961-2 (1963), 61 if.
Tumuli and cart-graves

A. ALFOLDI and A. RADNOTI, 'Zligelringe und Zierbeschlagevon romischen Jochenund
Kummetenaus Pannonien',Serta Hojjilleriana (Zagreb, 1940), 309 ff.
E. BONIS, Pol. Arch. ix, 1957, 76 ff.; xiv, 1962, 23 ff.
H. KERCHLER, 'Die romerzeitlichenBrandbestattungen unter Hligeln in Niederosterreich',
ArchaeologiaAustriaca, BeiheJtviii (Vienna, 1967).
R. MULLER, Arch. Ert. xcviii, 1971, 3 if.
s. PAHIC, Arh. Vestn. xi-xii, 1960-I, II6 if.; xix, 1968, 321 ff.
K. SAGI, Arch. Ert. 1943, 113 ff.; lxxviii, 1951,73 ff.
M. SEPER,ArheoloskiRadovii Raspraveii, 1962, 335 if.
Inhumation
L. BARKOCZI, Acta Ant. xiii, 1965, 215 ff.
L. BARKOCZI, KomaromMegyei MuzeumokKiizlemftryei i, 1968, 75 if.
A. SZ. BURGER, Acta Arch. xviii, 1966, 99 if.
A. SZ. BURGER, Pol. Arch. xix, 1968, 87 if.
F. FULEP, Arch. Ert.lxxxix, 1962, 23 if.; xcvi, 1969,3if.
V. LANYI, Acta Arch. xxiv, 1972, 53 if.
L. NAGY, 'Mumienbegrabnisse aus Aquincum', Diss. Pann. i, 4 (Budapest,1935).
R. NOLL, Barb-Pestschr.149 ff.
K. SZ. POCZY, Arch. Brt. xci, 1964, 176 ff.
E. SPAJIC, Os}. Zb. xi, 1967, 101 if.
E. B. v AGO, Alba Regiai, 1960,46if.
M. v ALTROVIC, Starinar ii, 1885, 33 ff.; ii/I, 1906, 128 ff.
L. ZOTOVIC, Lu] i, 1961, 171 ff.
ARCHAEOLOGICAL RESEARCH: BOOKS AND PERIODICALS

GEOGRAPHICALLY ARRANGED
PANNONIA
A. GRAF, 'Ubersichtder antiken Geographievon Pannonien',DiS!. Pann. i, 5 (Budapest,1936).
Tabula Imperii Romani, SheetL 34, Aquincum-Sarmizegethusa-Sirmium (Budapest,1968).
E. B. THOMAS, R VP.
423
SelectBibliography
Austria
Burgenl. Heimatb.
Fundberichteaus Osterreich
jahrbuchfur Altertumskunde
JOAl
Pro Austria Romana
RLiO
L. FRANZ and A. NEUMANN, Lexikon Ur- und FruhgeschichtlicherFundstattenOsterreichs(Vienna,
1965).
w. KUBITSCHEK, RO'merfunde.
w. KUBITSCHEK, '.Altere Berichte tiber den romischen Limes in Pannonien,i', Sitzungs-
berichte der OsterreichischenAkademie,Philosophisch-historischeKlasse ccix/1 (Vienna, 1929)'
G. PASCHER, 'RomischeSiedlungenund Strassen im LimesgebietzwischenEnns und Leitha',
RLiO xix (Vienna, 1949)'
B. SARIA, 'Der romischeGutshofvonWinden am See',BurgenliindischeForschungen,xiii (Eisen-
stadt, 1951).
H. VETTERS and H. MITSCHA-MAHRHEIM, Die RO'merzeit in Niedero'sterreich, Atlas von
Niedero'sterreich(Vienna, 1958).
Vindobona
A. NEUMANN, Vindobona. Die riimische Vergangenheit Wiens, Geschichte-Erforschung-Funde
(Vienna-Cologne-Graz,1972).
Carnuntum
Bericht des VereinsCarnuntum
Carn.-jb.
E. SWOBODA, Carnuntum.
Hungary
Acta Ant.
Acta Arch.
Alba Regia
Ant. Tan.
Arch. Ert.
Arrabona
Fol. Arch.
VMMK
K. BAKAY, N. KALICZ and K. SAG!, Veszpremmegye regeszetitopogrdft4ja, A keszthe!Jijdrds
(Budapest,I966). A devecseries sumegijdrds(Budapest,I970).
J. BANNER and I. JAKABFFY, ArchiiologischeBibliographie des Mittel-Donau-Beckens
(Budapest,
I954), I9J4-I9J9 (Budapest,I961), I96o-I966 (Budapest,1966).
407
SelectBibliography
I. ERI, M. KELEMEN, P. NEMETH and I. TORMA, Veszpremmegyeregeszetitopogrdjirija, A vesz-
premijdrds (Budapest,1969).
J. FITZ, A rrfmai kor Fg"fr megyeben(Szekesfehervar,1970).
F. FULEP and E. CSEREY, Nagytete'!JmiiemJekei(Budapest,1959).
Intercisa i-ii.
E. KOCZTUR, 'Somogy megye regeszetileletkatasztere',RegeszetiFiizetek ii, 13 (Budapest,
1964).
B. KUZSINSZKY, A Balaton kiir'!Jekenekarchaeologirija (Budapest,1920).
R. MULLER, Regeszeti terepbljdrdsokagikslji 'szegek'videken(Zalaegerszeg,1971).
M. WOSINSKY, Tolna vdrmegyetiirtenete az iiskortrfl a honfoglaldsigi-ii (Budapest,1896).
Savaria
Savaria, Bulletin der Museendes KomitatesVas
Vasi Szemle
T. P. BUOCZ, Savaria-SzombathelJ topogrdjirija (Szombathely,1968).
Scarbantia
Soproni Szemle
K. SZ. POCZY, Sopron rdmaikori emJekei(Budapest,1965).
Brigetio
BARKOCZI, Brigetio.
Aquincum
Bp. Miieml.
Bp. Reg.
L. NAGY, Az rfbudai 6keresztCt!Jcella trichora a Raktdr utedban(Budapest,1934).
L. NAG Y, AZ Eskii teri eriid, Pest vdros iise (Budapest,1946).
J. SZILAGYI, Aquincum(Budapest-Berlin,1956).
Gorsium
Alba Regia
J. FITZ, Gorsium (Szekesfehervar,1970).
Sopianae
JanusPannoniusMuzeumEvkiinyve
F. FULEP, Peesrrfmaikori emJekei(Budapest,1964).
425
SelectBibliography
Yugoslavia
Arch. lugosl.
Arh. Pregl.
LuJ
ZA
Slovenia
Arh. Vestn.
Glasnik MuzdskegaDrustva za Slovenijo
Situla
Varstvo Spomenikov
J. KLEMENC and B. SARIA, Archiiologische Karte von Jugoslawien, Blatt Pltij (Zagreb, 1936);
Blatt Rogatec(Zagreb, 1939)'
Emona
J. SASEL, Vodnik po Emoni (Ljubljana, 1955).
M. DETONI and T. KURENT, Modularna rekonstrukcijaEmone(Ljubljana, 1963).
Poetovio
M. ABRAMIC, Fuhrer durch Poetovio (Vienna, 1924).
J. KLEMENC, Pttijski grad v kasni antiki (Ljubljana, 1950).
Neviodunum
T. KNEZ, P. PETRU and s. SKALER, Neviodunum(Dolenjske Zalozbe, 1960).
Croatia
Arheolofki Radovii Rasprave
Od. Zb.
VHAD
Vjesnik arheolofkogmuzdau Zagrebu
J. KLEMENC and B. SARIA, ArchiiologischeKater von Jugoslawien,Blatt Zagreb (Zagreb, 1938).
B. VIKIC and M. GORENC, Prilog istraziva1!fa antickih nase!Jai putovau deverno-zapadnrjHrvatski
(Zagreb, 1969)'
Serbiaand V ojvodina
Radvrjv. muzo
Starinar
ZNMB
D. and M. GARASANIN, Nalazista.
426
SelectBibliography
MOESIA SUPERIOR
Yugoslavia(seeabove)
Serbia(seealso above)
ArheoloIki nalaziIta i spomeniciu Srbiji, ii, Centraljna Srbija, Gradja x. ArheoloIkog lnstituta
(Belgrade,1956).
Stm'eKulture u Djerdapu-AnciennesCultures du Djerdap (Belgrade,1969).
Kosovo and Metohija

GodiI1ljak Muzo/a Kosova i Metohije
E. CERSKOV, Rimljani na Kosovui Metohiji (Belgrade,1970).
E. CERSKOV, Municipium DD, Socanica(Pristina-Belgrade,1970).
Macedonia
ZA
Zbornik, Recueildes Travaux, Publicationsdu MuseeArcheologiquede Skoplje
N. VULIC, Nekolikopita1lja iZ anticke istorije naIe zemije (Belgrade,1961),
Bulgaria
Arheologija
lBAl
s. GEORGIJEVAand v. VELKOV, Bibliograftja na balgarskataarheologija I879-I9JJ (Sofia, 1957).
BARBARIAN LANDS BEYOND THE FRONTIER
Czechoslovakia
Arch. RoZhl.
Pam. Arch.
Slov. Arch.
Hungary (seealso above)

DebreceniDeri MuzeumEvkijnyve
Egri MuzeumEVkb'nyve
MFME
Miskolci Herman Otto MuzeumEVkiJnyve
M. PARDUCZ, Denkmiiler der Sarmatenzeit
in Ungarn i-iii (Budapest,1941-51).
427
I Index of Deities, Peoplesand Persons
Achilles, on grave reliefs, 262 Andia, personalname, 65, 254

Aconii, at Andautonia,136 Andinus, Upper Moesian god, 254
Aecorna(Aequorna),goddess,74, 182 Andinus, personalname, 65, 254
Aelii, as decurionsof municipia, 144 f. Andio, personalname, 65,254
Aelius P. f. Aelianus, P., 143 Andizetes,Pannoniantribe, 14, 53,61; civitas of,
Aristides, 70, 109, III, 181 66, 137
Caesar,L., 102 Anemius, bishop of Sirmium, 331
Catus, 37,43,66 Annaei, at Neviodunum,136
Dasius, P., 143 ANSA, legendon coins, 56
P. lib. Menander,P., 134 Antigonus Doson, king of Macedonia,9
Plato, P., 134 Antinous, shrine, 134
Triccianus, 245, 248 Antonii, customs-farmers,226
Aemilius Aemilianus, 205 f. Antoninus Pius, emperor, building under, 162,
Aeneas,on grave-reliefs,261 f. 168; citizenship grants under, 141; Danube
Aequitius, magistermilitum, 292 fr. frontier under, 103 f.; milestones, 107;
Aequorna,seeAecorna Mithraism under, 181; provincial organiza-
Aesculapius,statueof, 326 tion under, 70
Africans, 226; as immigrants to Pannonia,230 Antonius of Lirina, St, 353
Agatharchides,26 Antonius, M., 21 fr.
Aggeus, Christian martyr, 326 Naso,praefectusof Colapiani, 69, 135
Agilimundus, king of Quadi, 288 Peculiaris,Titius, customs-farmer,226
Agrippa, M. Vipsanius, 34, 53, 68 Q.f. Cl. Firmus, 121
Akichorius, Celtic prince, 25 Apollodorus of Damascus,bridge built by, 98
Alans,34° Aponius Saturninus,M., governor of Moesia, 42
Alaric, leader of Visigoths, 345, 347 f. Appian, 13, 15,21 f., 27,44,109;IIlyrike, 31,200
Alatheus,Gothic leader, 340 fr., 345, 349 f., 352 f. Appiarenses,in Pliny, 68
Alexanderthe Great, 5 Appius Claudius Pulcher, 17
Amantini, Pannonian tribe, 14, 34, 53 fr., 6o; Aquileians, in municipia, 136
Roman organizationof, 66, 70, II4 Arabiates,civitas, 66; organizedby Romans,54 f.
Amantius, bishop of Jovia, 341 f., 347 f., 353 Araharius, king of Sarmatians,288
Ambigatus, Celtic king, 5 Ardiaei, 13
Ambrose, St, 342, 344 f. Argaragantes,279 f., 288 f.
Amicenses,279 Ariogaesus,king of Quadi, 189 f.
Ammianus Marcellinus, 103, 272, 279, 286 fr., Arius, followers of, 330
293 fr., 302, 310, 342, 348 Arminius, 40
Anartii, Celtic tribe, 18 f., 35, 58 Athanarich, 342
Anbo, Celtic personalname, 61 Athaulf, leaderof Visigoths, 347 f.
429
Index oj Deities, Peoplesand Persons
Atilius Saturninus,L., 78 Brigoi, Thracianpeople,9
Atta Bataionisfilius, negotiator, 79, 152 Britons, in auxiliary units, 154
Atta Nivionis filius, hoard of, 175 Burebista,king of Dacia, 17 fr., 25, 27, 32, 34 f.,
Attalus, king of Marcomanni,206, 345 37,61
Attila, leaderof Huns, 349 fr., 357 Buri, defeat of, 191
Aturo, Celtic name, 59,61 Busturo, Celtic name, 59
Augustus,emperor,98; at Aquileia, 34; campaign
against lapodes, 21 fr., 27, 30, 32 f., 114; CaecinaSeverus,army commander,38,43
citizenship grants, 57; frontier policy, 34 fr., Caesar,seelulius
74,89,185;populationand settlementunder, Caesernii,Aquileian family, 77, 120 fr., 136
53 fr., 66 fr.; provincial garrisonunder, 73 CaeserniusZosimus, 121, 125
Aulus Gellius, II9 Caesii, Aquileian family, 77
Auluzon, Thracianname, 65 Caligula, see Gaius
Aurelian, emperor,3I I ; deathof, 266; proclaimed CalpetanusRantius Quirinalis, c., governor of
emperor, 210; religion under, 259; reorgan- Pannonia,80 f.
ization of Dacia by, 2I I, 273 f.; wars of, Candidus,cognomen,248
211 Canii, Aquileian family, 77,120, 122
Aurelii, native, 214 Capenii, at Neviodunum, 136
Aurelius, ]'vI., name of decurions,223 Caracalla,emperor,201, 218; building, 194; pro-
Constantius,altar dedicatedby, 298 vincial administration, 142, 215, 221, 223;
Victor, 100, 242 religious policies, 258; wars, 198 f.
Aureolus, army commander,206 f., 210 Carinus, emperor,267
Ausonius, 291 Carni,32
Autariatae,Illyrian tribe, 2 fr., 9, 17, 25, 27 Carnuntina,freedwomanof veteran,77
Avars, invasion of, 352 f., 358 Carpi, 198, 203, 209, 272 f., 306
Avidius Cassius,rebellion of, 184, 190 Carus,emperor,267
Avitus, emperor,350 f. Cassius,decurionof Mogentiana,144
Azali, Pannoniantribe, 54 fr., 58 f., 61; civitas of CassiusDio, 23 f., 37 f., 84, 184, 186, 190 fr., 198,
66,137,145,155 201, 299
Longinus, c., campaignof, 32
BaebiusAtticus, c., primipilus, 69 Catali, 59
Ballomarius, king of Marcomanni,186 Catari, Pannoniantribe, 59,61; civitas, 66
Banadaspus,king of lazyges,190 Catualda,Germanking, 40
Barbii, Aquileian family, 77 Caupianus,princepsof Boii, 137
Barbius L. lib. Nymphodotus,130 Celegeri,68, 223
Bargates,freedwomanof, 227 Celts, as part of pre-Romanpopulation I fr., 3I,
BassaeusRufus, praetorianprefect, 191 55, 59 fr., 358; conqueredby Dacians,37, 61;
Bastarnae,allies of Dacians, 23 f., 35; as mer- in army, 200; in local aristocracy,248; settle-
cenaries, 23; campaign against Dardanians, ment of, 55, 58; tumuli of, 151
10,18,26;submissionto Rome, 41 Cerethrius,Celtic leader, 7, 25
Bato, chieftain of Breuci, 34, 38 f., 56 Chatti, 84
Bato, chieftain of Daesidiates,37 fr. Ciliunus, Celtic name, 59,61
Bato, Pannonianname, 59 Cimbri, migration, 13 f.
Belgites, civitas, 66 Cittu, native personalname, 65
Belgius (Bolgius), Celtic leader, 7, 25 Claudian, 344 fr., 351
Bellovesus,Celtic prince, 5, 25 Claudius,emperor,55, 74, I I 5; citizenshipgrants,
Bessus,from Scupi, 65, 249 79, 114; colonization under, 50, 76, 78 f.;
Bithynia Severa,grave-inscriptionof 239 frontier policy, 40, 47 fr.; provincial ad-
Bitus, Thracianname, 65 ministration,69 f.; provincial garrisonunder,
Boii, besiegeNoreia, 17; civitasperegrinaof, 66, 69, 4 2, 48, 51, 70, 155
137; driven out by Marcomanni, 35; in ClaudiusII (Gothicus),emperor,210, 249
Roman period, 53, 55; in S.W. Slovakia, Fronto, M., provincial governor, 187, 204
57 f.; land laid wasteby Dacians,6I ; villas in Marinus Pacatianus,Ti., army commander,204
territory of, 169, 173, 176, 244 Pompeianus,army commander,188
Bolgius, see Belgius. Priscus,Tib., praefectusalae, 114 f.
BONOSA, nameon gold pin, 353 Clodius Albinus, 201
Brennus,Celtic leader, 7, 9, 25 CocceiusCaupianus,M., tombstoneof, 135
Breuci, civitas, 51, 53, 66, 137, 145; na~es of, Colapiani,civitas, 51,66,69,II4; namesof, 60 f.;
59 fr.; Pannoniantribe, 14, 53; recrUitment Pannoniantribe, 14; organizedby Romans,
from, 39, 5I f., 155; revolt of, 34, 38 f., 54 f. 54 f .
Breucus,Pannonianname, 59 Commodus,emperor,defenceof frontier, 194 fr.,
430
Index of Deities, Peoplesand Persons
zz8; foreign policy, 193; rebellion against, Dio Cassius,see Cassius
zoo; wars, 191 f., 197 Chrysostom,91
Constans,emperor,z85, Z91, 306, 308 Diocletian, emperor, 3lZ; abdication of, Z76;
Constantia,daughterof Constantine,z85 formation of Tetrarchy, z67; frontier policy,
ConstantineI, emperor, Z7, z67, z85, Z91, 30Z, z69 ff., z80, z85; palaceof, 306; persecution
307, 3II, 313 f.; Christianity under, 323 f., of Christians, Z59, 326; provincial re-
3z8, 332; estates,306; limes under, z8z, z85; organization,zz3, Z73 ff.; wars, z68 f.
prosperityunder, 3IZ; provincial administra- Diodorus,7
tion, Z73 ff.; wars, Z77 f., z88 f. Dioscuri, Z54
ConstantiusI, emperor,z67 Ditybistos, 358
bishop of Lauriacum, 353 Diurpaneus,king of Dacians,8z
II, emperor,z86 ff., Z93 f., 329 f. Dolens, Thracianname, 65
Cornacates,civitas of, 5I, 66, 137; organizedby DOMISA, legendon coin, 56
Romans,53 f.; Pannoniantribe, 14 Domitian, emperor, colony of, II 6; death, 9 I ;
Cornelius Felix Plotianus, governor of Lower frontier under, 47, 89; provincial garrison
Pannonia,197 under, 81 ff., 85 f., lZ0, lZ4; trade under,
Fuscus,procuratorof Pannonia,41; praetorian lZ0, lZZ, lZ4, lZ9; wars, 8z ff., 10Z, lZ9
prefect, 8 Zf. Domitius Ahenobarbus,governor of Illyricum,
Scipio Asiagenus,L., 15, 18 f., 32 35
Cosmius,customsofficial, zz8 Zmaragdus,C, decurion of Carnuntum, 141,
Cotini (Kytnoi), incorporatedinto empire, 57 f., ZZ7
188,190,19z, 199, z09, zzz, Z48, Z7Z; in pre- Domnus, bishop of Sirmium, 330
Roman period, 19, 35; language of, 59 f.; Donatus,priest, 327
pay tribute to Vannius, 40 Drigissa, Thracianname, 65
Cotiso, Dacian king, Z 3f. Drusus, 39 f., 44
Crassus,see Licinius Dubius, scriba, 144
Critasirus, king of Taurisci and Boii, 19, z7 DVTEVTA, on coin, 56
Dacians,z7, 47; civitas, 66, 68; early Romancam- Ennodius,344

paign against,ZI ff., 3Z, 35 ff., 4Z; invasionof Epona,worship of, z53 f.
Moesia, 41 ff.; languageof, 4; recruitmentof Eppii, from Gaul, 136
praetorians from, zoo; settlement of free Eravisci, 152; civitas, 66,137,141f., 151,155ff.;
Dacians, 191; wars against Celts, 17ff., Z5, coins, 56; names, 59 ff.; Pannoniantribe, 5,
35,37,55,61;wars againstRome, 8z ff., 91, 54 f .
94 f., 100, IZ9 Eusebius,bishop of Cibalae, 325
Daesidiates,rebellion of, 37 ff. Eutherius,bishop of Sirmium, 330
Daizo, Thracianname, 65
Dalmatians,4, 5z, 117, z17 Farnobius,Gothic leader, 339 f.
Danuvius, god, 18z, Z31 Firmidii, at Neviodunum,136
Dardanians, 34, Z40; civitas, 68 f., 145, zz3; Flavii, native, 135 f., 138
exports of, Z46; in pre-Roman period, 5, Flavius Biturix, T., 135
9 ff., z5 ff., 153; names,65, Z54; pacification, Bonio, T., 135
Z3 f.; recruitmentfrom, Z36; wars of, 15 ff., Cobromarus,T., 135
ZI Lupus, senator,348
Dardanoi,9 Proculus,T., tombstoneof, 134 f.
Dardanos,z74 Samio, T., 135
Dases,Pannonianname, 59 Florus, Christian martyr, 326
Dasmenus,Pannonianleader, 38 historian, 36, 39, 185
Pannonianname, 59 princepsof Eravisci, 137
Das(s)ius,Illyrian name, 65 FonteiusAgrippa, governorof Moesia,4z, 80, 83
Dea Syria, Z 5 8 Fortunatus,Christian martyr, 327, 333
Decebalus,king of Dacia, 8z ff., 86, 95 Fragiledus,z88
Decius, Traianus,emperor,z03 ff., z44 f. Frigeridus, campaign in Illyricum, 339 f.; tile-
Decoratus,cognomen,z48 stampsof, Z91, Z93
Deianira, on mosaic, Z 3 7 Fritigern, Gothic leader, 339 f., 34z
Demetrius,king of Macedonia,9 Fritigil, queenof Marcomanni, 345
(Demetrios,Dmitrij) deaconof Sirmium, 327 Fronto, ZOI
Dentheletae,Z3 f., 34 Fufius Germinus,army commander,zz
Diana, goddess,Z31, Z51 ff. FunisulanusVettonianus,L., governorof Moesia
Diegis,84 Superior, 8z
Dimenses,in Pliny, 68 Furii, estateof, 138, Z41; freedmanof, z63
431
Furius Alcimus, freedman,241, 263 Hercuniates,civitas, 66, 145, 155,222;organized
Furtius, king of Quadi, 189 by Romans,54f.; origin, 17,63
Hermes,Christian martyr, 326
Gabinius, king of Quadi, 294 Hermunduri,35 if.
Gaiobomarus,king of Quadi, 199 Hermyle, Christian martyr, 328
Gaius, Christian martyr, 326 HerodesAtticus, 244
Gaius (Caligula), emperor,citizenship grants, 57; Herodian, 197 if., 299
trade under, 78 Herodotus,2
Galerius,emperor,277, 326; as ruler of Illyricum, Homer, I f., 17
267 if., 276; economic policy, 265 f., 272; Horace, 23
improvementwork, 298, 306; palaceof, 302; Huns, 329, 345; eifect on frontier, 295 f., 342,
provincial reorganization,273; residenceat 346; empire of, 339, 349 if., 357; in alliance
Sirmium, 3I 2 with Alans and Goths, 340 f.
Galla Cnodavi f., 124 Hyginus, II9
Gallienus,emperor,21 I, 266, 307, 345; Christian-
ity under, 325 f.; evacuationof Daciaby, 209, Iallius Bassus, governor of Pannonia Superior,
259; wars, 105, 203, 205 if., 210, 264 f., 186
286 Iapodes,21 f., 24, 32
Gallonius, scriba, 144 Iasi, civitas, 5I, 66, 137; names of, 60 f.; Pan-
Gallus, Caesar,286 nonian tribe, 14, 54, 5 8
Gauls, settlers in towns, 76, 117, 125, 136; in Iazyges, 93, 195 f., 209; alliance with Quadi, 40;
legions, 154 client-treaties with, 41 f., 91, 100, 102;
Generidus,army commander,348 f. peace-conditions,101, 192 f., 197; settlement
Genius Ciniaemus,Pannoniangod, 253 in Hungarianplain, 19 if., 37, 39, 56, 86 if.,
Gepidae,209 if., 267, 272, 277 f., 322, 345, 358 94 f.; wars, 84 if., 94 f., 99 f., 102, 189 if.,
Germanicus,38 194; see also Sarmatians
Germans, 35, 59 f.; Caracalla'spolicy towards, Jerome,339, 341, 347, 35 8
198 f.; client-state system, 42, 95, 102; Jews, 228
Domitian's policy, 89; frontiers with, 92 f., Illyrians, as part of pre-Romanpopulation, I if.,
284; in auxilia, 154; pacification of, 189, 31, 60, 65, 248, 358; as representativesof
192f.; settlements,58; wars, 83 if., 103, 187, Illyricum, 200; god of, 250; in auxilia, 154
191, 194, 202 Illyriciani, 290, 302; compositionof, 242 f., 245;
Germinius, bishop of Sirmium, 330 headquartersat Sirmium, 245, 254, 276 f.,
Getae, 1,4; Roman campaignagainst, 23; settle- 323; political power of, 201 f., 204 f., 207 f.,
ment in Moesia, 37,66 210 f., 244, 249, 253, 265 if., 277
Gordian III, emperor,203, 208, 218, 225 Ingenuus,206
Goths, Christianity among, 353; effect on limes, Jonah,on sarcophagus,334
202 f., 209, 211 f., 267, 272; migration, 342, Jordanes,279
349, 352; plundering raids, 339 if., 349; Jovian, emperor,267, 290
receptio, 296, 339, 351 f.; rulers of Danube Iphigenia, on grave reliefs, 262
lands, 350; treatieswith, 280, 339, 341; wars, IRAVSCI, on coins, 56
204 f., 210 f., 264, 278 f., 289, 321 f. Irenaeus,bishop of Sirmium, 327
Gratian, emperor, 294, 324; campaignsagainst Isis, 254; shrine of, 244
Goths, 339 if. Istri, 32
Greeks,I f., 26 f., 138; in Christiancommunities, Italians, at Scupi, 117; in commerce, 120, 124,
32 3 130, 136, 223; in legions, 154, 230; settle-
ment by, 13 I, 169, 178; stonemasons,179;
Hadrian,emperor,82, 98,127,134,162,175,179; villa-owners, 124
as governor of Pannonia Inferior, 92, 95; Italicus, prince of Quadi, 41, 83 f.
evacuationof Dacia by, 101; frontier policy, Iucundusprinc. Azalus, 70
104 if.; mining under, 133; provincial garri- Julian, emperor,290, 324, 330
son under, 99,155,181;urbanization,II8 f., Julii, customs-farmers,226
136 if., 151, 157, 161, 168, 218 f., 315; wars, Iulius, decurionat Mursella, 144
100, 102 f. Caesar,c., 18 if.; Commentaries,17, 63
Haterius Nepos, Ti., governor of Pannonia Magimarus,C(aius), 57
Superior, 103 QuadratusBassus,c., governorof Dacia, 100
Hecataeus,2 Juppiter Dolichenus, cult of, 181, 255 f., 258 f.,
Hector, on grave reliefs, 262 325
Helvetii, 13 Optimus Maximus, altars, 125, 182, 250 f.;
Heraclianus,praetorianprefect, 2I 0 votive inscriptions, 324; worship, 253
Hercules,on mosaic, 237 (Thracian), 254
432.
Justin, emperor,350, 358 Maroboduus,king of Marcomanni,37, 40, 43
Justin, epitome of PompeiusTrogus, 7 Mars, altar to, 125; cult of, 254; on grave reliefs,
Justinian, emperor, 27, II6, 224, 276, 280, 303, 262
350 ff., 356; birthplace, 214 f. Latobius, worship, 253
Iuventii, at Andautonia,136 Martin, of Bracara,St, 353
Iuventius Primigenius,M., dedicationby, 136 Matres Pannoniorum,234, 250
Matumarus,137
Kaulikoi,2 Maximian, emperor,267 f., 302
K ytnoi, see Cotini Maximilla, Christian martyr, 327
Maximinus, 294
Lares, worship of, 252 Thrax, emperor,202 f.
Latobici, civitas, 5I, 66, 136; in pre-Roman Memmius Regulus, governor of Achaea, Mace-
period, 5 3 migration
; of, 17, 63 ; tombstones, donia and Moesia, 44
151 Mercator, vi/ieus, 100
Laurus, Christian martyr, 326 Merobaudes,295
Lentulus, campaignon Danube,36 f., 39, 43 Mestrianus,native name, 65
Leonianus,35 3 Mestrius, native name,65
Leonnorius,Celtic leader, 9, 25 Minitra, native god, 253
Liber Pater,altar to, 298 Mithras, altar to, 276; cult of, 181, 255 ff., 324;
Liburni,4 symbols, 254; votive tablets, 246
Liccaius, Pannonianname, 59 Mithridates, 15, 18,23
Licco, Pannonianname, 59 Moesi (Moesoi, Mysoi), I f., 9, 17, 27, 68 f.
Licinius, emperor, 277, 323, 328; baths of, 312; Montanus,Christian martyr, 327
deeennalia,300 Mucatral, Thracianname, 65
Crassus,:1\1., 21, 23 f., 32 f., 44 Mucianus, see Licinius
Mucianus,42, 8I Mucco, Thracianname, 65
Limigantes, 279, 288 ff. Mysoi, see Moesi
LivY,5
Lombards,186, 192,353 Naevius Campanus,L., 86
Lucian, 103 PrimigeniusNaristus, 129
Lugii,84 Nanea,Thracianname, 65
Lusii, at Andautonia,136 Naristae,189 f., 192
Lutarius, Celtic leader, 9, 25 Nemesis,cult of, 254; shrine to, 162
Neptune,altars to, 125, 136
Macedonians, 2; at Scupi, II 7; Bastarnae as Nero, emperor,50 f., II5, 135; citizenshipgrants,
mercenariesof, 23; in legions, 155; relations I 14; provincial garrison under, 48; settle-
with Dardanians,9 f., 26 ment of barbarians,41, 66, 68
Macrianus,207 Nerva, emperor, citizenship grants, 135, 137;
Macrinius Vindex, praetorianprefect, 188 milestonesof, 107; war, 85, 89, 102
Magnentius,usurper,285 f., 330 Nessus,on mosaic, 237
MagnusMaximus, emperor, 342 f. Nicetas, bishop of Remesiana,351
Maidoi, 12, 15 Noah, portrait of, 334
Maiorianus,dux et praesesof Valeria, 340 Noricans, 24, II7, 154
Marcellianus,dux Va/eriae, 294 Numerian, emperor,267
Marcellinus comes,349 f. Nutrices of Poetovio, 250
Marcianus,army commander,210 Nymphidius Sabinus,c., 50
Marcii, Italian family, 121 f., 136
Marcius Turbo, Q., 100, 204 Obii, 186
Marcomanni,4°,71,89,342;migration of, 35 ff., Obulenses,native community, 68
43,56; relationswith Caracalla,198 f.; settle- OctavianusCaesar,see Augustus
ment within empire, 206 f., 209, 345; wars, Oetenses,native community, 68
37, 84, 101, 186 f., 189 ff., 203, 268, 272 Oppius Sabinus,governorof Moesia, 82 f.
:l\1arcus Aurelius, emperor, 141, 209, 244, 253; Opponii, Aquileian family, 170
frontier policy, 80 ff., 195 f., 202, 227 f., 230; OptatianusPorfyrius, 278
mining under, 63, 131 ff., 195, 213, 216 f.; Orosius, 344
municipalization by, 213 ff., 240, 313; re- Oseriates,civitas, 66,145,155; namesof, 61; orga-
cruitmentunder, 154, 183 f., 195,236;settle- nized by Romans,53,55;Pannoniantribe, 14
ment of Cotini, 19, 57, 190, 199, 209, 248, Osi, in pre-Romanperiod, 19, 35; languageof,
272; tradeand industry under, 129, 178, 245; 59 f.; pay tribute to Vannius,40; remainout-
villas, etc. under, 169, 175; wars, 183 ff., 197, side empire, 57
200,247,264,286 OsiuS,330
433
Pacatianus,206 RAVIZ, on coins, 56
Pacatus,potter, 177 f. Regalianus,usurper,206 ff., 264
Paionians,65 Resatus,potter, 177
PalpelliusHister, governorof Pannonia,41 Rhea Sylvia, on grave reliefs, 262
Pan, on pottery, 177 Rhinelanders,125, 130 f.
Parthians,201 f., 228 Richomeres,army commander,339
Paul, St, in painting, 334 Rider-gods,Danubian,246, 254, 324
Paulinus,345 Romula, mother of Galerius, 302
Paulus,354 Romulus,Christian martyr, 327
Pelagonians,65 Roumanians,356
Peregrinus,sutor caligarius, 129 Roxolani, 94 f., 191, 193; migration of, 2°9;wars,
Perseus,king of Macedonia,1 ° 41 f., 100
Persians,202 Rubrius Gallus, governorof Moesia, 42, 80
Pertinax, emperor,188 Runo, 288
PescenniusNiger, 200 f., 228 Rustii, at Neviodunum,136
Peter, St, in painting, 334
Petronii, Aquileian family, 77 Sabinii, customs-farmers,226
PetroniusL. lib. Licco, L., 124 Saco,Celtic name, 59
Philip II, king of Macedonia,9 Saphrac,Alannic leader, 340 ff., 345, 349 f., 352 f.
V, king of Macedonia,9 f. Sarapis,cult of, 258
Philippus, emperor,203 ff. Sarmatians,47, 59 f., 89, 93, 186, 293, 296, 322,
Photeinos,bishop of Sirmium, 330 345, 358; trade with, 245; treaties, 89, 9 1,
Phrygians,9 271, 288 f.; wars, 36 ff., 42, 84, 95, 99, 103,
Picenses,Picensii, Moesiantribe, 68 187, 192 f., 197, 202 f., 205 f., 209 ff., 264,
part of Limigantes, 279 267 ff., 276 ff., 286 f., 290 f., 294 f., 310, 342;
Pinnes,leaderof Pannonianrebellion, 39 see also Iazyges
Pipa, daughterof Attalus, 206 Savarina,freedwoman,77
Plautius Silvanus, crushes Pannonianrebellion, Savus,god, 136, 182
38 f. Scerviaedus,Illyrian name, 65
Aelianus, governorof Moesia, 41, 66 ff. Sitaes(Sitae filius), 65
Pliny the elder, 19, 37, 53, 59 f., 66 ff., 74 ff., 1I2 Scipio, see Cornelius
Pollio, lector at Cibalae, 325, 327 Scordisci (Scordistae), 25 ff., 30, 53; allies of
Polybius, 12 f., 28 Rome, 34, 39; campaignof Tiberius, 23 f.,
PompeiusSextus,governorof Macedonia,38 34; Celtic tribe, 5 ff.; civitas, 66, 137; de-
PompeiusTrogus, 5, 7 cline, 15 ff.; hegemony,9 ff., 202; namesof,
Pontii, estateof, 138, 241 60, 65 f.
PoppaeusSabinus,governorof Moesia, 33,44 Scorilo Ressatilib., 129
Posidonius,I f., 13 ff., 17 f., 28 Scribonius Curio, c., Macedoniancampaign of,
Priam, on grave reliefs, 262 2, 17 f., 32
Priscus, 35 I, 357 Scythians,68
Probus, emperor, economic policy, 265 ff., 272, SecconiusPaternus,c., 125
298 f.; estatesof, 244 Sedatus,god, 253
praetorianprefect of Illyricum, 294 f., 31o SeptimiusSeverus,emperor,and Danubianarmy,
Procopius,131 ff., 350 f. 197 f., 200 f., 204, 217; immigration into
Pseudo-Augustine,344 towns under, 228; municipal reorganization
Ptolemy, 18, 21, 53, 57ff., 66ff., 110 by, 140, 214 f., 218 ff., 225 f.; religion under,
232 ff., 256, 258
Quadi, 35, 37, 60, 89, 342, 345; fort in territory of, Houseof, 110,126,148,193,198,248,25 1,256,
293 f.; under Vannius, 40 f., 57 f.; wars, 263,265,3°7;client-kingdomsunder, 198 f.,
83 f., 101, 103, 186 ff., 199, 203, 268, 286 ff., 2°9; Danubianarmy of, 200 f.; immigration
291,294ff. under, 227 ff.; mining under, 134, 216;
Quadriviae,see Silvanae prosperity under, 230 ff.; provincial ad-
QuartusAdnamatif., 151 ministration, I I 9, 2I 7 ff., 240; trade, 247
Quintilius Condianus,army commander,188 Serapilli, Pannoniantribe, 54; civitas, 66, 145, 155
Maximus, army commander,188 Serretes,Pannoniantribe, 54; civitas, 66
Quintillus, 210 Severi, see SeptimiusSeverus,Houseof
Quirinus, bishop of Siscia, 326 ff., 348, 353; Severinus,St, 357
basilica of, 3I 3, 332 SeverusAlexander,emperor,202, 216
Sido, prince of Quadi, 41, 83 f.
Radagaisus,barbarianleader, 347 ff., 353 SidoniusApollinaris, 344, 351
Rausimodus,Sarmatianleader, 277 f. Sigovesus,Celtic prince, 5, 25
434
Sigynnoi,2 Tigidius Perennis,praetorianprefect, 197, 199 f.
Silius Nerva, P., 24 Timachi, Moesiantribe, 68
Silvanae(Quadriviae),deities, 25 I f. Timens, native personalname, 65
Silvanus,worship of, 250 ff., 254, 299 Titus, emperor,82
Augustus,cult of, 250 f. TraianusDecius, see Decius
domesticus,252 Trajan, emperor,98, 102, Il2, 125, 181, 340, 354;
Messor, 252 citizenship grants by, 141, 143, 145; estab-
Silvestris, 252 lishment of Dacia as province, 21, 60, 86 ff.,
Sindoi,2 91 f., 95, 100, 227; foundation of colonies,
Sinna, Thracianname, 65 Il6, Il8 f., 157, 173; limes under, 47, 89,
Sisciani, Pannoniantribe, 5 I I06£., lo9ff.;miningunder,131 ff., 153,213;
Sita, Thracianname, 65 provincial administration, 68, 92, 134 ff.;
Sol, 254; cult of, 259 provincial garrison under, 80, 85 ff., 92, 96,
Spaniards,226; at Scupi, II7; in army, 154 99, 124, 129; recruitment,155 ff.
StephanosByzantios, 3 I 3 TrebonianusGallus, emperor,203, 205
Stilicho, 344 f., 347 Triballi, civitas 68 f., in pre-Romanperiod, 2 ff.,
Strabo, 13 ff., 18 f., 27 f., 61, 66 9, 17, 25; pacification, 23 f., 27
Stratonice,Christian martyr,328 Triccus, Celtic name, 248
Strubilo, Germanname, 59 Tricornenses,reorganizedMoesian tribe, 68
slave, 129, Tricornensii, Moesiantribe, 68
Suebi, 84, 89, 99, 192, 199, 20 5, 20 9 Trojans, 27
Suetonius,39 Trotedii, at Neviodunum,136
Sulla, 15 Tudrus, dynasty of, 89
Sulpicia Dryantilla, denarii of, 206 Germanname, 59
Synerotas,Christian martyr, 327 f., 333 slave, 129
Syrians, 201; at Scupi, 117; immigration of, Tullius Germinus,governorof Moesia, 68
227 ff.; in auxilia, 154; in legion, 155
Ulpia Andia, I 3 8
Tacfarinas,44 Ulpian, jurist, 241
Tacitus, emperor,266 Ulpii, native, 137 f., 145
historian, 19,40,43,51,57,59 f., 71, 83, 89, 91 Ulpius Andinus, 138
Taifali, 279, 289; invade Dacia Ripensis,340, 342 Fronto Aemilianus, M., 143
Tampius Flavianus,governorof Pannonia,41 f., Timentis f. Maximus, M., 138
197 Ursacius,bishop of Singidunum,329 f., 332
TarruntenusPaternus,190 Usafer, barbarianking, 288
Taurisci, 32, 35, 53, 55, 57; early hegemony,13, Uttiedius Celer, Sextus,77
18 f., 25
Tectosages,settlementof, 63 Valao, king of Naristae,189
Teitia, Celtic name, 61 Valens, bishop of Mursa, 329 f., 332
Terentianus,dux of Valeria, 291 emperor,291, 339 f.
TerentiusVarro Lucullus, M., 17 Valentinian,emperor,203, 267, 290 f., 295 f., 302,
Tertullian, 185 322,331; decline of towns under, 310, 312;
Tettius Iulianus, army commander,83 fort-building by, 89, 291 ff.
Teutanus,worship of, 253 Valentinian II, emperor,295
Thana,altar to, 25 I Valeria, daughterof Diocletian, 273
Theodosius,emperor, 308, 346, 351; as dux of Valerian, emperor,205, 325 f.
Moesia Prima, 294; monetary circulation Valerianus,son of Gallienus, 205 f.
under, 343; wars, 340 f., 343 ff. Valerius Dalmatius,governorof LugdunensisIII,
Theopompus,25 342
Thracians,65, 68; as part of pre-Romanpopula- Maximianus,M., 245
tion, Iff., 358; immigration, 230; in auxilia, Messalla,governorof Pannonia,37 f., 154
I 54; in mining area,2 I 7 in
; praetorianguard, Seutesdomo Bessus,L., 125
200 Vandals, 192 f., 198 f., 209 f., 267, 272, 278, 322,
Tiberii Iulii, tombstonesof, 120 345; admission to Dacia, 189; attacks on
Tiberius, emperor, 40, 50 f., 71, 114, 154; cam- limes, 186,203,2II, 267; migration, 347, 349
paign against Scordisci, 23 f., 30, 34; cliff Vannius, king of Quadi, 40 f., 57 f., 102
road of, 45 ff.; foundation of colony by, 74; Varciani, civitas, 5I, 66; namesof, 6o; organized
pacification of Pannonia,34, 37 ff., 54 ff., 61, by Romans,53, 55
66, 152; provincial administration,49, 68 f., Velleius Paterculus,36 ff., 71, 76
74, Il5, 167, 173; provincial garrisonunder, Venantius,altar dedicatedby, 298
36, 43 ff.; trade, 78 Veneti, 2,4
435
Vergil,262 Visumar, king of Vandals, 279
Verus, emperor,187 Vitali anus,commanderof Pannoniantroops, 341
Vespasian, emperor, 41 f., 49, II6, 135, 154; Vitellius, emperor,41
foundation of colonies by, 112 f., 1 3°; Vitrodurus, king of Quadi, 288
frontier under, 80 ff. Vlachs, 3 5 5f.
Vetranio, 285 f. Volcacius Primus, L., cohort prefect, 69, 135
Vettius Sabinianus,governor of Dacia, 191 Vulcan, on pottery, 177
Vibius Logus Hermundurus,slave, 129
Victohali, 279, 288 f. Zanticus,king of Iazyges,190
Victoria, on intaglio, 72 Zenobia,2II
Victorianus, bishop, 323 Zeus, 1
Victorinus, bishop of Poetovio, 326 Zimarchos, 358
Vidasus, god, 251 Zinafer, 288
Vinicius, M., army commander,35 ff., 56 f. Zizais, king of Argaragantes,288 f.
Visigoths, 345 Zosimus, 348
436
2 Index of Places
Achaea,44 decline of, 31of.; excavation,159 If.; foreign

Actium, battle of, 21, 23, 33 settlers, 120 If., 124 If., 228; fort (auxiliary),
Acumincum (Slankamen),279 50, 73, 80 f., IIO, 143; fortress (legionary),
Ad Flexum (Magyar6var),fort, 88 85 f., 92, 99,101 f., 105, 110 f., IF, 183, 198;
Ad Militare (Kiskoszeg-Batina),fort, 88 grantedrank of colonia, 2I 8; headquartersof
Adony, see Vetus Salina governor, 94, III, 237; houses,161, 164 If.,
Adrianople, 340, 344 237,239,300,310,317; industry, 122, 161 f.,
Aegean,the, 212 177 f., 181, 237; municipium, 139 If., 157;
Aelianum, 225 roads, lI8, 122 f., 137, 144, IF, 162, 244;
Africa, 44, 194 f., 338 slave-trade, 129; walls, 16I, 168; and see
Albertfalva, fort, 88, 1°5;civil settlement,238 Budapest
Alessio, see Lissus Arabo (Raba,Raab), R., 54, 121, 144, 198, 354
Alexandria, 109, 245 Arear, 354; and see Ratiaria
Alisca (Ocseny),fort, 106 Arelate (ArIes), mint, 343
Alma Mons (FruskaGora), 38, 246, 298 Argentares,354
Alutus (Olt), R., 96, valley, 97 Arrabona (Raab, Gyor), 144, 354; fort, 49 f., 69,
Amida, 287 81,88,135,186
Andautonia,municipium, II 5, 136, 152, 155, 213 Asia Minor, 1,7,9,25,27, 210, 238, 328; cults
Angros, R., 2 from, 258; immigrants from, II7, 228;
Annamatia,fort, 105 legionariesfrom, 154 f.; trade with, 245, 322
Antioch, 340; immigrantsfrom, 81, 141,227 Astibus (Bregalnica),R., 13
Apollonia, coins of, 30 Aszar, hoard, 175
Apulum, 218 Athens, 210
Aquae, Chora of, 133, 224, 3°3; see at, 329 Atrans, 286
Aquae, fort, see Prahovo Augst, silver-hoard,337
Aquae Balizae (Aquae Balissae),(Daruvar), 143 Aureliana (= municipium Aurelianum?), 13 I,
Aquae Iasae,144 133,224f., 273, 276, 310
Aquileia, 22, 33 f., 124, 181, 187, 228, 250, 299, Aureus Mons, mountain, 298
330, 341, 348, 353; coins from, 321; im- Aureus Mons (Seona),municipium, 224 f., 298
migrantsfrom, 76 If., 120, 130; mosaics,302, Axius (Vardar), R., 9, 44; as route, 322; site of
3I 3, 337; road to Danube, 44, 7I ; trading Scupi on, 115 If.; valley, 12, 214, 340
centre, 28, 31 f., 123, 246
Aquincum, 88, 175, 182, 196, 230, 262, 269, 277, Bakony, mountains,144
290, 292 f., 295, 337; amphitheatre(civil), Balaca,villa, 124, 171, 243, 307
162; (military), 169; Christianity, 333, 336; Balaton, lake, see Pelso
citizenship grants at, 57 f., 145; civil settle- Banat, the, 19, 97, 101
ments, 126, 140 If., 226; cults at, 256 If.; BanatskaPalanka,see Lederata
437
Index oj Places
Bardovce,II6 126, 168; municipium, 139 f.; promotion to
Bassiana,240; municipium, 143, 152, 159,166 If.; colony, 218; roads,44,5°,71,76f., 115, 152;
promotedto colony, 225 f.; walls of, 3I 2 settlementof veterans,77; slave-trade,129
Bathinus (Bosna),R., 56 CastraConstantia,see Ulcisia Castra
Batusa,224 CastraMartis, 340
Bela Palanka,see Remesiana CastraRegina (Regensburg),184
Belgrade (Beograd), 63, 68, 223; and see Singi- Celeia (Celje = Zilli), 77
dunum <,;:elamantia,see IZa-Leanyvar
Berkasovo,337 Cerevic, II4
Beroia, 117 Cezava,see Novae
Berytos, 117 Ciabrus (Cibrica), R., 68
Berzovia, 101 Cibalae (Vinkovci), 203, 267, 277, 286, 290, 315;
Black Sea, I, 82, 322, 342 Christianity at, 325 If., 329; Cotini settledin
Bohemia(Boiohaemum),14,71 territory, 199, 248, 272; municipium, 143,
Bononia, 269, 278, 326, 354 152; promotedto colony, 225 f.; road, 244
Bor, 131 Cibrica, R., see Ciabrus
Bosnia, 12, 30 Colapis (Kulpa), R., 12,22,54,II4; valley, 51
Botivo, see Jovia Cologne, 125, 130
Bozevac,224 Constantinople, 277, 287, 290, 340; and see
Brae, 312 ByZantium
Bracara,35 3 Cornacum(Sotin), 53, 143
Braunsberg,73 Cremona,83
Brazda lui Novae, seeDevil's Dyke Cserszegtomaj,cemetery,124
Bregalnica,R., see Astibus Csopak, 336
Brestovik, burial vault, 238 Cuppae(Golubac),47 f., 96
Brigetio (Szony), 106, 144, 186, 2°3, 230, 288, Cynoscephalae,battle of, 9
292, 294 f., 302, 325; civil settlements,140;
cults at, 256 If.; foreign settlers, 120 If., 124,
Dacia, 4, 18, 21, 84, 86, 91, 183, 265, 279, 340,
126, 228 f., 260; fort (auxiliary), 49 f., 88: 342; barbarian attacks on, 203, 205, 208 f.;
fortress (legionary), 85, 88 f., IIO; garrison, clearing of barbariansfrom, 187 f.; coloniza-
92, 99; joins Pa=oniaInferior, 198; muni- tion of, 130, 227; conquest by Rome, 89;
cipium, 22 I; promotion to colony, 225 f.; establishmentof province, 94 f., 97 f., 100,
roads, 122, 162, 3°5;tombstonereliefs, 122 130; evacuation,101, 209 if., 267 f., 280, 3II,
Britain, 107 354; flight from, 60; garrison, 88, 92, 95 If.,
Brongos,R., see Morava 198,203;reliefs of Rider-godin, 254; roads,
Brundisium, 23, 98 97 f., 100f., II8, 138, 214; Sarmatians
Brza Palanka,see Egeta grantedaccess,191, 193, 197,209;settlement
Budalia, 244 of barbarians,189, 191 f.; trade, 130
Budapest, native coins minted at, 56; native Aurelian's province, 267,272 If., 3II, 354
oppidumat, 18,72,173;and see Aquincum Dacia Mediterranea, establishment, 274 If.;
Bulgaria, 55, 277 Gothic raids on, 349; peacein, 296; persecu-
Byzantium, 350; road to Singidunum,196,213f., tion of Christians, 326; seesin, 329; settle-
246, 26o; and see Constantinople ment of barbariansin, 341
Dacia Ripensis, 302 f., 349, 354; establishment,
Caesariana,244, 307 274; failure to export, 321; invasion of, 340,
Campania,72 342; limes, 280, 285, 296; peace in, 296;
Campona (Budapest-Nagyteteny), coin-hoard, persecution of Christians, 326; sculpture,
279; fort, 101, 1°5,277,282If. 3II; sees in, 329; settlement of Goths in,
Cannabiaca(Klosterneuburg),88 339,343,352;Vandal migration from, 347
CariCin Grad, 215 Dalj, see Teutoburgium
Carnuntum, 45, 56, 73, 104 f., 110, 188, 200 f., Dalmatia,34, 38 f., 65,82, 139,223,348;minesin,
206, 262, 293, 295; amphitheatre(civil), 164, 131,208;native population, 4,13 f., 27, 55;
227; (military), 169; auxilia at, 50; Boian provincial administration, 134 f., 140; re-
aristocracy at, 151 f., 238, 243; civil settle- cruitmentin, 154, 242; settlersfrom, 113, 134
ments, 126 If., 140; conferenceof emperors, Damascus,98
276, 302, 324; cults at, 181, 256, 324; decay Danube,passim
of, 310; entrep6t at, 120; excavations,159, Dard (....), municipium, 133; and see Soeanica
162 If., 237, 310; foreign settlers, 120, 124, Dardania,354; emperorfrom, 249; establishment
227; fortress (legionary),48,69,71,80f., 85, of province, 273 if.; exports, 299, 322;
284, 317; garrison, 40, 43, 92, 99; head- in pre-Roman period, 9 If., 15; legionary
quartersof governor, 94, 186; market place, fortress in, 69; part of East Roman empire,
438
Index of Places
35 0; peacein, 296; persecutionof Christians, Germany, 59, 92; exports, IZ2; lower Germany,
326; personalnamesin, 65; raidedby Goths, 140
349; recruitmentin, 154, 196; seesin, 329; Gigen, see Oescus
Theodosiusin, 340, 343 Galle, 305
Daruvar, see Aquae Balizae Golubac,see Cuppae
Delphi, 7, 15 Gomolava,oppidum, 73
Deutschkreuz,villa, 299 Gornja Klisura (upper Djerdap), 45, 96
Devil's Dyke, 271, 279 ff. Gorsium (Tac), fort, 5I, 104; civil settlement,
Dierna (Orsova),45 ff., 97, 130 173 ff., 315 ff.; imperial cult at, 175
Dimitrovica, see Szavaszentdemeter Gracanica,145
Djerdap (Iron Gates),1,7,36,48,81,88,96,107, Granua(Gran, Garam,Hron), R., 188, 354
28o; canal and road through, 45 ff., 109 f.; Greece,7
see also Gornja Klisura Gyal6ka,fort, 104; soldiers'tombstones,50
Donji Milanovac, see Taliata Gyor, seeArrabona
Donnerskirchen,villa, 336
Drave, (Drava, Dravus), R., and valley, 12 ff., 27, Haemus,Mt, 340, 349
34,38 f., 48,5°,55f., 92,1°5,151,169,272, Halimba, 175
343, 350, 354; boundary of Pannonia, 33; Hebrus (Marica), R., 13
civitates on, 5I, 66; finds in area, 71 f.; Herculia, 307
frontier betweenPannoniaPrima and Savia, Hercynia silva, 5, 18, 54,61 ff.
273; municipalizationof, II8, 143,222,309, Hetenypuszta(Jovia), fortification, 305 ff.
341, 348; personalnamesin, 61; routes, II8, Hiulca palus (Volcaepaludes),38, 272
322; settlementofjoederati,342; tribal names, Horreum Margi, 5I; developmentof, 2 I 5; in-
54; villas, 173 scriptionsat, 23 I, 236, 241; municipalization
Drobeta (Turnu Severin),98, 130 of, 196, 214 f., 225; recruitmentat, 236, 241;
Dubravica, 216 see at, 328 f.; walls, 215, 312
Dunaalmas,burgus, 107 Hron, R., see Granua
Dunakam16d,see Lussonium Hrtkovci, 73
Dunaszekcso,see Lugio Hungarianplain, 84; peoplesof, 2, 19, 36 f., 39,
Dyrrhachium,coins of, 30 88,94,245,32 2,342,347,349,358
Hungary, 2, 7, 28, 49, 106, 269, 277
Egeta (Brza Palanka), 96 ff.; Dolichenum at,
25 6 Ibar, R., 2,131; valley, 133,223
Egypt, 245 Ig,76
Eisenstadt,villa, 171 Illyria, 5, 13
Elbe, R., 35 Illyricum, 23, 214, 268, 276 f., 287, 290, 295, 299,
Emona (Ljubljana), 30, 33, 51, 59, 61, 70, 308; 313,33°,343,357; constitutionof province,
cults at, 182; foundation of colony, 74 ff.; 34 f., 53, 56; division of, 39, 42; in third
group of citizens at Savaria, 74, 76, 78; century, 200 ff., 208, 211; pacification of,
Italian traders at, 71, 74, IZO; legionary 39 f.; refugeesfrom, 347 f.; road building,
fortress (?), 43; road, 286; stonemasonsat, 44 f.; threatenedby barbarians,339 ff.
50; street-system,74, 159; transferred to Intercisa(Dunaujvaros),fort, 49, 101, 105 f., 195,
Italy, 94; walls, 74, 178 227; housesat, 238; inscriptions, 197, 222,
Erzgebirge,131 230; Syrian settlersat, 227 f., 258, 260
Esztergom,see Solva Iron Gates,see Djerdap
Isker (Oescus),valley, 17, 44, 2IZ
Fenekpuszta,120 ff., 124; Christian basilica, 336; Istria, 2, 59; invasion of, 24, 34
fortification, 302 ff., 319, 353 Italy, 4 f., 14 f., 18, 24, 32 ff., 41 f., 44, 55 f., 59,
Ferto, lake, 58, 61, 244 83, 94, 180, 194, 197, 199 f., 205 f., 208 f.,
Fischamend,burgus, 107 2II, 277, 286, 308, 330, 343 f., 347, 350;
Florentiana,354 emigration to, 327, 348, 353; invasion of by
Flortin, 354 barbarians,103, 187; routes to, 10, 32, II3,
Fossis(Jarak),road station, 298 202; settlementof barbariansin, 189, 278,
Frigidus, R., battle of, 345 28o; settlers from, 76; stonemasonsfrom,
179,181;trade,traders,31, II9 f., 123, 129 f.,
Galatia, 15, 26 152, 299, 342
°
Gamzigrad,imperial hunting-castle,3 5 f., 337 IZa-Leanyvar(Celamantia),fort, IIO f.
Garam,R., see Granua
Gaul, 56,61, 100, 294, 298, 321, 338, 347, 353, Jablanica,valley, 214
356; exports, 122 Jarak,see Fossis
Gellert, Mount, oppidum, 72 f., 86, IIO, 173 Jovia, fortification, see Hetenypuszta
439
Index oj Places
Jovia (Botivo), municipium, 222, 225, 309; seeat, Margum (Orasje), 267, 278; cives Romaniat, 131;
222, 329, 341 f., 347 developmentof, 215 f.; promotion to muni-
Jugoslavia,277 cipium, 196, 214; see at, 329, 349, 352
Julian Alps, 31 f., 4°,187,345,347 Mariniana, 240
JustinianaPrima, 214 Maros (Marisia, Marisos), R., 100 f., 279
Matrica (Szazhalombatta),civil settlement, 173,
KaliSte, 224 238,315 ff.; fort, 49,106f.
Kasan,gorge, 45 ff., 96 ff., I09 Mattersdorf,fort, 104; tombstone,50
Keszthely-Ujmajor,122 Mecsekmountains,152, 223
Kiskoszeg-Batina,see Ad Militare Mediana,villa, 286, 291, 302, 324, 337
Kladovo, see Pontes Metohija, 9, 63 ff., 138, 145, 234
Klagenfurt, 3I Metulum,22
Klisura, Serbian,44, 214 Milan, 330, 345
Klosterneuburg,see Cannabiaca Misenum, fleet of, 52, I I 3
Kopaonik mountains,131 ff. Mitrovica, see Szavaszentdemeter
Korkoras (Krka), R., 135, 166 Mlava, R., 131; valley, 96, 215, 223 f.
Kornye, fortification, 305 Moesia Inferior, 68, 208, 2II f., 274; creation of,
Kosmaj, Mount, inscriptions, 23 I; mmmg area, 82; garrison, 83,97; settlementof Goths in,
63, 133 f., 188, 195,213,216f., 224 f., 300 296, 339, 352; wars, 203 ff.
Kosovo and Metohija (Kosmet), 138 Prima, 289, 294, 296, 343; establishmentof,
Kosovo Polje, 9, 98,115 f., 131, 138 273 ff.; in fifth century, 349; limes, 279, 280,
Kostol, palace,224, 300 285 f., 346; seesin, 329
Krajina (Serbian),224 Secunda,establishmentof, 274; settlementof
Kraljevo, 223 Goths in, 339
Kulpa, R., see Colapis Mogentiana (Tuskevar), municipium, 145, 152,
Kusadak,statecameo,300, 337 166
Mons Claudius, 38
Laibach marshes,40 Morava (Brongos), R., 13; river transport, 131,
Latina, border-station,260 2I 6; native population of, 2, 19, 65; road
Lauriacum, 184, 353 through, 51, II6, 213 ff., 347; tombstones,
Lederata(BanatskaPalanka),97 236,241,249;valley of, 44,195,223,268
Leitha, district, names in, 59; tombstones,135, Moravia, 89, 173
150 f., 179; villas, 135, 164, 169, 173, 299 Moravica, R., 223
Leopoldsberg,73 Mosel, R., 291
LepenacR., II6 f. Mukacevo, 18
Leshja, see Lissus Murocincta, see Parndorf
Leskovac,214 f. Mursa (Osijek), 53, 85, 206, 222, 273, 285, 315,
Lezoux, pottery, 178 341; foundation of colony, 119; oriental
Lipljan, 354 settlement,228; pottery finds, 177 f.; roads,
Lirina, 353 120, 122 ff., 137, 143; see at, 329 f.; settle-
Lissus (Leshja, Alessio), 98, 214 ment of Cotini around, 199, 248, 272
Ljubljana, see Emona Mursella (Kisarpas,on R. Raba), 285, 354; fortifi-
Locice, fortress, 184 cation at, 305 f.; municipium, 144 f., 152, 222
Lucus Vocontiorum, 117 Mursella (Petrijevci, on R. Drave), fortress (?),
LugdunensisIII, 342 85; municipium (?), 222, 225
Lugdunum(Lyon), 234; mint of, 343 Mutatio Sarmatarum,268
Lugio (Dunaszekcso),fort, IOI, 105, IIO, 268 Mysia, I
Lussonium(Dunakomlod),fort, 50, 88, IIO
Nadldki Hrib, fort, 5I
Macedonia, 2, 15, 23, 32, 38,43 f., 65, 68, 117, Nagyteteny,see Campona
2I 0, 26o; before Roman conquest,7 ff.; in- Naissus(Nis), 2, 195,210,267,277,285 f., 290 f.,
vasions of, 17 f., 21, 24, 34; Moesian army- 302, 354; auxiliary soldier at, 44, 51;
commandsubordinateto governor of, 36 f., Christian community, 326, 332; conquered
43; recruitment in, 154, 242; settlementof by AttiIa, 350; fortress (?), 44, 48, 5I, 85;
Sarmatiansin, 28o; trade, 299, 322 foundation of municipium, 98, 196, 2I4f.;
Maestriana,240 grave-chapel, 3 I3, 334 f.; imperial work-
Magdalensberg,3I shops,337; in fifth century, 351; inscriptions,
Magyarovar,see Ad Flexum 231, 236, 241, 315; late Roman cemetery,
Malata (Banostor),fort, 50, 88 313,332,335;prosperity, 3II, 313 f.; roads,
Marcal, R., 354 97,206,212,214,225,246; seeat, 329; trans-
Marcomannia,89; projectedprovinceof, 184, 193 ferred to Dacia Mediterranea,274 ff.
440
Index oj Places
Natiso, R., 3 I Po valley pottery, 88, III, 122, 129, 177
Nauportus (Vrhnika), 43; Italian traders at, 30, Poetovio, 41, 222, 235, 245, 309, 341, 343, 345;
33,76 bishop of, 326; ceded to Noricum, 273;
NeusiedlerSee,see Ferta,lake citizenship grants, 137; colony, 94, I I 8 f.;
Neviodunum (Drnovo), municipium, I 15, 166, foreign settlers,71, 15 I; fortress, 43, 48, 71,
2 I3; grave-inscriptionsin area, 235 85, 118 f.; garrison, 43, 48, 69, 77, 88, 209,
Nicaea, council of, 328 259; industry, 122, 124; inscriptions,43,23°,
Nis, see Naissus 244; Mithraeum, 181,2°9,256,259; Nutrices
Nisava, R., 214, 3 I 3 of, 250; palaceat, 302; roads, 71, 76, 122 f.,
N6gnidverace,322 137,144,152,286;veteransettlement,II8 f.,
Noreia,I7 173: and see Ptuj
Noricum, 17, 35, 40, 56, 76 f., 104, 183 f., 273, Poigen, 106
348, 353, 357; barbarian invasion of, 103; Pola,loo
exports, 322; frontier, 83; iron, 3I; mines, Polgardi, 337
IF; namesin, 57, 59, 61 Pontes(Kladovo), 98
Noricum_Ripense,ducate,348 Pontus,IoO
Novae (Cezava),fort, 48, 96 Poreeka,R., 45, 96
Novi Sisak, 114 Potaissa,187
Nyergesujfalu,burgusnear, 107 PraesidiumDasmeni,5 I
PraesidiumPompeii, 5I
Ocseny,see Alisca Prahovo(Aquae?), fort, 89
bdenburg,see Scarbantia Pristina, 131 fr., 145
Oescus,R., see Isker Ptuj (= Pettau), II8, 354; and see Poetovio
Oescus(Gigen), centreof Triballi, 2, 25; fortress,
43,45,69,92,2IZ; road to, 44; stonemasons' Quadrata(BaratfOldpuszta),fort, 107
workshop, 18o
Olt, R., seeAlutus Raab,see Arrabona
Oltenia, plain of, 95, 99 Raab,or Raba, R., see Arabo
Opitergium (Oderzo), 187 Raetia, 267, 348; trade, 322
Orasje, see Margum Rakos, R., 110
Orsova,see Dierna Ratiaria (Arear), centreof Moesi, 68; colony, 98,
brvenyes,villa, 336 118 f., 226; conquered by Attila, 350;
Osonibus,57 fortress, 82 f., 85, II8 f., 2IZ; imperial estate,
244; in fifth century, 351 f.; inscriptions,230,
Paeonia,12 f. 3 I5, 35 1; names of legionaries from, 65;
Paestum,II 3 persecutionof Christians, 326; recruitment
PannoniaPrima, 313, 343, 345, 348; constitution from, 154, 157, 24of.; roads, 97f., II8f.,
of, 273; ducate,348; limes, 284, 291; persecu- 214, 225; sarcophagi,237, 246; sculptures,
tion of Christiansin, 326, 328; seesin, 329 18o; seatof administrationof Dacia Ripensis,
Secunda,279, 289, 298, 348; campaignagainst 276; seeat, 329, 352; settlementof easterners,
Sarmatiansfrom, 287; establishmentof, 273; 130; tombstoneof auxiliary, 5I; transferred
limes, 286, 346; persecution of Christians, to Aurelian's Dacia, 273
328; seesin, 329; trade, 299, 322 Ravenna,40, 347; Court of, 346; fleet of, 52, II 2 f.
Parisos,R., see Tisza Ravna, see Timacum minus
Parndorf, villa (= villa Murocincta?), 135, 171, Regensburg,see CastraRegina
173,238,244, 295, 299, 302 Remesiana(Bela Palanka), Christian basilica at,
Patka,pottery, 176 335; grave-chambers,3 I5; seat of concilium
Pees,see Sopianae provinciae, 214; see at, 214, 310, 329, 335,
Pek (Pincus), R., 68, 131; mining near, 96, 351 f.; villa, 3°°;walls, 312
13 1 Rgotina, 354
Pelso(Balaton),lake, 57 fr., 66,120,122,144,222, Rheinzabern,pottery, 178
303, 305 f.; canal, 272, 298; decline in area Rhine, R., 84, 202, 269; as linguistic boundary,
round, 123 f.; grave-inscriptionsnorth of, 6o; civil settlementsat fortresses,139; forts,
235 f.; villas north of, 169, 173, 243 f., 299, 107, IZO; frontier, 35,49, 185; troop move-
33 6 ments, 81, 85, 92, 124
Pergamum,26 Rhineland,influence on funerary sculptures,125;
Pettau,see Ptuj trade and traders, IZ9 f., 148
PiazzaArmerina, villa, 303 Rhodopemountains,344
Pincum (Veliko Gradiste),fort, 68, 96, 279 Rimini, 330
Pincus,R., see Pek Ritopek, see Tricornium
Placentia,II 7 Romuliana,palaceof Galerius, 302
441
Index oj Places
Romulianum,birthplace of Galerius, 267 aries from, 65; recruitment, 154, 157, 240;
Roumania,356 roads, 210, 214; sculpture, 180; seat of ad-
ministration of Dardania, 276; see at, 329,
Saale,17 352; street-system,II 6; theatre, I 16, 181;
Sagvar,seeTricciana tombstones,82, II6 f., 314 f.
Sala (Zala), R., 120, 144, 303, 354 Scythia Minor, 280
Sala, municipium, see Salla Segestike,see Siscia
Salamanca,tombstones,61 Seona,see Aureus Mons
Salla (Sala),municipium, 144 f. Serbia, 2, 7,131,215
Salona(Split), 38; emigrationto, 354; palace,306; Serdica (Sophia), 44, 286; bellum Serdicense,2°7;
roads, 44 f.; veteransfrom, II 7 Constantineat, 277 f.; council of, 328 ff.; seat
Sarmatia,282; building of forts in, 269, 293; proof administrationof Dacia Mediterranea,276
jectedprovince, 184, 193; underConstantine, Singidunum, 133, 166, 177, 224, 238, 267, 290,
278 f. 300; auxiliary troops in area of, 188, 195;
Sarmizegethusa, 82 civil settlement,131, 140, 223; conqueredby
Sar Planina,see ScardusMons Huns, 350; foreign settlers,130; fortress, 82,
Sarszentmikl6s,wagon-grave,148 85, 91, 96, 99, 101, 140, 223; inscriptions,
Savaria (Szombathely),II8, 144,278,285,353; 130,230,232;mines in neighbourhood,133,
capitolium, 78; cemeteries, 79, 313, 335; 188, 195,213;Mithraism, 258; municipium,
Christianity, 328 f.; citizens of Emonaat, 74, 223; persecution of Christians at, 327 f.;
76,78; colony, 76 ff., 94,135,162,3°8; cults promotionto colony, 225 f.; roads, 143, 196,
at, 244, 256; decay, 310; execution of 206, 2I 3 ff., 246, 26o; sculpture, 18o; see at,
Quirinus, 326 ff.; foreign settlers,77,120ff., 329 f.; soldiers' families, 241; tombstones,
151, 228; imports, 124; inscriptions, 74, 78, 130,188; and see Belgrade
23°,235,3°8,315,333; Kalendarium Septim- Si6, canal, 272
ianum, 219; mosaics,244, 313; new citizens, Sirmium (SremskaMitrovica), 73, 166, 226, 265 f.,
II3, 137; palace, 302, 313, 332; relics of 3II, 354, 357; attachedto Moesia Superior,
Quirinus, 313, 332,348; roads,50, 115, 122£., 82, 86; attacked by Pannonians, 38, 43;
137, 144, 244, 303 ff.; seat of administration barbarianattackon, 294; baths,3I 2; cededto
of Pannonia Prima, 273, 3II, 314 ; silvae East Rome, 349 f., 352; cemeteries,II 3 f.,
dominicae, 306; statue to Capitoline Triad, 3I 2 f., 327, 33 3, 33 5, 348; Christianity at,
181,324;stonemasons,50; street-system,78, 259,323, 325 f., 330 f.; citizenship grants at,
159; tombstones,122, 124, 151 f., 181,244; 137; coin-finds in region, 28 ff.; colony, 86,
walls, 78, 310, 312 93 f., 112 ff., 130, 135, 137; conqueredby
Save,R. and valley, 12, 18, 22, 43, 48, 92, 99, 104, Huns, 350; court workshops, 337; cults at,
129, 169, 178, 216, 223, 244, 286, 347, 350, 254,258; decline in importance, 276 f.,
354; bridge-building, 44; citizenship grants 290 f.; emigration from, 327, 354; estates,
in, 136; civitates, 5I, 54, 66; founding of II 5, 244; excavations,I I 3 f., 312 f.; fortress
towns, 113 ff., II7, 120, 136, 143,213,222; near, 85 f.; garrison in area, 43, 82, 86;
pacification of, 30, 32 ff., 55, 57; peregrini, harbour, 114; hippodrome, 312; horreum,
58 f.; roads to, 44 f.; settlements,74 ff., I 13; 312; houses,312; immigration to, 228, 323,
trade and communicationsroute, 28, 72, 98, 325; imperial residence,188, 190, 202, 210,
II3, 115, 120, 130, 136 f., 322; tribes, 7, 10, 268,272,277,285ff., 289 ff., 295, 312 f., 323,
13 ff., 17 f., 24,27f., 32 ff., 38 f., 55, 57,61 33 I, 340 f.; importancein third century, 202,
Savia, administratorof res privata in, 306; con- 204, 206 f., 210, 267, 285; inscriptions, 230,
stitution of province, 273; in fifth century, 313, 315, 323, 351 f.; mint, 343; names at,
350; MagnusMaximus in, 342 f.; persecution 59; palace, 302, 312; persecutionof Chris-
of Christiansin, 326, 328; seesin, 329; settle- tians, 326 ff.; Probus's economic reforms,
ment of barbarians,341 f. 266, 272, 298 f.; roads, 45, 206, 244, 286,
Scarbantia(Sopron, Odenburg),flight from, 348, 3°3; samianfinds at, II4; seatof administra-
353; foreign settlers, 15 I, 173; inscriptions, tion of PannoniaSecunda,273, 314; seat of
78,235; municipium, 115, 135, 159, 166 ff.; commanderof army of PannoniaInferior and
oppidum, 73 f., 76, 78, I I 5; paganism in Moesia Superior, 204; see at, 323, 327,
fourth century, 324; stonemasons,50; villas, 329 ff., 349, 352; shops, 312; territory, 114;
173; walls, 166, 312 theatre, 294, 310; transferred to Pannonia
ScardusMons (Sar Planina), 9, 98, 195, 241 Inferior, 93 f.; walls, II 3 f., 294, 3ro, 3I 2,
Scupi (Skoplje, Dskiib), 138, 145, 241, 354; 348
Bessusfrom, 249; cemeteries,II6; colony, Sisak, 354; and see Siscia
II6 f., II9; earthquakes,356; estates, 307; Siscia (Segestike,Sisak), 27, 272,354;besiegedby
fortress (?), 44, 48, 51; houses, II6; in- Romans, 12 f., 32; capturedby Magnentius,
scriptions, II9, 230 f., 260; namesof legion- 286; by MagnusMaximus, 343; by Octavian,
442
Index of Places
22 f., 34, 114; centreof Colapiani, 51; colony Thessalonica,327, 340, 343; road, 210, 214
at, 94,112if., 130, 135, 137; headquartersof Thrace, 4, 7, 24, 139,212,23°,260, 277, 340 f.;
Tiberius, 38, 51, 114; inscriptions, 222, 230, barbarian attacks on, 210 f.; cults from,
3I 5 ; measuresof Septimius Severusat, 2I 9, 253 f.; grave-chambers, 238; legionaries
244; mint, 208, 343; possiblefortress,43, 48, from, 154; miners from, 134; provincial re-
71; recruitment at, 155; road, 143; seat of organization, 274; recruitment from, 230,
administration of Savia, 273, 3II; see (of 242; settlementof barbarians,280, 339, 341;
Quirinus), 326 f., 329, 332, 348; 'Siscia' pot- tribes of, 9, 13,26, 34
tery, 177 f.; traders'tombstone,7I Timacum minus (Ravna), civil settlement, 98,
Skoplje, see Scupi 225; fort, 51, 81, 97, 195; inscriptions, 230;
Slankamen,seeAcumincum sculptures, 18o; tombstones of soldiers'
Slovakia,Celtsin, 14, 19,28, 57; Iron Age houses, families, 241
173; Romanbuildings, 89 Timok, R., as waterway, IF; cult in, 254; route
Socanica,223; and see Dard ( ....), municipium through, 98, 214; tribes, 17, 68; valley, 44,
Solva (Esztergom),293 51,81,97,133,214,347
Somly6vasarhely,144 Timok, Black, R., 214, 224
Sophia,see Serdica Timok, White, R., 214, 224
Sopianae(Pecs), 294; Christian monuments,3I 3, Tisza (Parisos),R., 37, 279; boundaryof Dacia,
329, 333 if.; foreign settlers, 120 f., 137, 19,21,86,95;plain of, conqueredby Iazyges,
222 f.; municipium, 222 f.; portrait-bust 95,99
from, 337; roads, 303, 3°5; seat of adminis- Tokod, 144,317
tration of Valeria, 222 f., 273, 310 f., 314; Tolna, 105
settlementof Carpi around, 306 Toplica, R., 2, 131; valley, 214
Sopron,see Scarbantia Transdierna(Tekija), 97 f.
Sotin, see Cornacum Transylvania,100
Spain, 140, 338, 353, 356; influence on personal Traprain, silver-hoard,337
namesand tombstones,61 Treviri (Trier), mint, 343
Split, see Salona Tricciana (Sagvar), cemetery, 334; fortification,
SpodnjaHajdina, II8 305 f.; municipium (?), 222; potteryfind, 122
SremskaMitrovica, II4, 312; and see Sirmium Tricornium (Ritopek), 68, 177
Srijem (Syrmien, Szoreny,Szeremseg),354 Trier, 295, 303, 340; and see Treviri
StaraPlanina,mountains,83 Trieste, see Tergeste
Stari Sisak, I 14 Troy, 9; on grave reliefs, 262
Stobi, 9; veteransfrom, II 7 Turnu Severin,seeDrobeta
Stojnik, fort, 195 Tusculum,inscription, 35
Stridon, 341, 344
Strymon (Struma),R., 13,44 Ulaka, auxiliary fort, 51, 104; native hill-fort, 51
Stimegcsehi,144 Ulcisia Castra(Szentendre),civil settlement,238;
Syria, cults from, 258; immigration from, 227 if. ; renamed Castra Constantia, 282; Syrian
imports from, 245; Pannoniansserving in settlersat, 228
vexillation in, 155; troopsfrom (on Danube), Ulpianum (Ulpiana), 224, 354; cemeteries,314,
81, 194 f., 227 332; citizenship grants, 138, 145; cults at,
Syrmien, see Srijem 254; grave-chapel,334f.; inscriptions, 230,
Szavaszentdemeter Mitrovica (Dimitrovica), 354; 236, 240, 249; latifundium, 241; literacy at,
and see Sirmium 263; municipium, 98, 133, 145, 240; recruit-
Szazhalombatta, seeMatrica ment, 154, 157,236; roads, 210, 214; sculp-
Szeged,entrepot,100 f.; inscription, I 10 tures, 18o; see at, 329; transferred to
Szekszard,sarcophagus,334 Dardania,276; walls, 168, 312
Szeremseg,see Srijem Dsktib. see Scupi
Szob, cemetery,7
Sz6ny, see Brigetio Valeria, 287 f., 290, 294, 334, 337, 339 f., 342 f.,
Szoreny,see Srijem 353; Christianity in, 326, 329; establishment
of, 273; improvementsof Galerius, 298; in-
Taban,industrial centre, 72 f. vaded by barbarians,277, 286; seat of ad-
Tac, see Gorsium ministration of, 223, 273, 310; stampedtiles
Taliata (Donji Milanovac), 45; fort, 48, 81, 96; of duces, 291 if.; troops in, 346; villas, 300,
road, 97 f. 322
Tapae,battle of, 83 f. Varceia, 53
Tekija, see Transdierna Vardar, R., see Axius
Tergeste(Trieste), 33, 59, 71 Veliko Gradiste,see Pincum
Teutoburgium(Dalj), fort, 50, 88 Venice, 228
443
Index oj Places
Vetus Salina (Adony), cives Romani consistentes, settlements,140; fortress, 85, 88 f., 92, 99;
143; fort, 49, 88, 101, 106 municipium, 221: and see Vienna
Vicus lovista, 143 Virunum, sculptorsfrom, 181
Vi din, 354 Visegnid, burgus, 293; fort, 282
Vienna (Wien), I, 7, 28, 354 Volcae paludes,see Hiulca palus
Viminacium (Kostolac), 201, 223, 285; canabae,
168, 218, 237; civil settlements, 130, 140, Wallersdorf, 73
218; communications,97 f., 215; conquered Westerndorf,pottery, 178, 245
by Huns, 350; foreign settlers at, 130, 141; Wien, see Vienna
fortress, 48, 73, 82 f., 85, 91, 9 6, 99, 21 5; Winden am See,villa, 164, 173
grave-monuments, 237; inscriptions, 73,
129 f., 218, 230, 232, 315, 351; meeting-place Zagreb,I36
of emperors,341; mint, 208; Mithraism, 258; Zajeear,224, 300
municipium, 139 if., 145, 218; promotion to Zala, R., see Sala
colony, 218, 225 f.; recruitment at, 157; Zalalov6, 144
sculpture, 129 f., 180 f., 237, 246, 261, 315; Zalavar, tombstones,120 f.
seatof administrationof Moesia Prima, 276, ZapadnaMorava, 2
3 II; of Moesia Superior, 215; see at, 329, Zegovac,mount, 133, 145
352; soldiers' families, 241; stonemasons Zemplin,I8
from, 216; trader from, 357; Trajan's head- Zgornja, I 18
quarters,92 Zilli, seeCeleia
Vindobona, 73, 191, 354; burning of, 345; civil Zlokucane, II6
444
3 Index of Subjects
administration,provincial, 49, 53 If., 69 f., 134 If., arithmetic, teachingof, 262

242, 273, 3II, 330, 343; seats of, 215, 218, arlit C•.•), functionary of civitas, 141 f.
223,273,276, 307 f., 312 If., 319 artefacts,Celtic, 7
agriculture, 299, 321; god of, 252 aryballos,on tombstones,148
AgrippinensesTransalpini, 162 atrium, classical,166, 171
aisles, of cathedralcrypt, 33 5 attributio, 142
Ala II Asturum, 50, 81 augurs, altar set up by, 142; grave of, 325;
VespasianaDardanorum,52, 154 Pannonian,234, 250
Frontoniana,155 Augustales,II7, 216, 222
I Hispanorum,50, 80 aulae, of palaces,302, 313 f., 332
I AugustaItyraeorum, 50, 154 auxiliary troops,administrationof civitatesby, 49,
I Noricorum, prefect of, 135 69; diplomas of, 58; in Moesia Superiorand
I Pannoniorum,50 Pannonia,43 f., 47 If., 79, 81, 85 If., 96 f., 99,
Scubulorum,51, 8I 104, 120, 130, 152, 186, 188; occupying
alue, on Danubefrontier, 49, 81, 120; in Moesia Dacia, 95 f.; protectingmines, 133, 195,217;
Superior, 9I ; prefectsof, I 14 f., 150; raising recruitment, 37, 39, 5I f., 57, 142, 154 If.,
of, 39, 52 195 f., 236, 239 If., 307, 346; Thracian, 38;
alleys, at Aquincum, 161 veterans,73, 104, II7, 142, 159, 175; see also
altars, 136, 182,231, 249 If., 256, 298, 3II, 336; ala; cohors; forts
at Aquincum, 86, 125, 141 f., 162; Brigetio,
50; Carnuntum, 276, 324; Naissus, 215; badge, brooches as, 320; lance as, 148; and see
provincial, of Lower Pannonia, 175, of lituus
Upper Moesia, 214; Singidunum, 130 baililfs, see viii ci
altar-slabs,marble, 336 barbarians,settlementof within empire, 41 , 57,
amber,trade in, 322 66, 138, 184 f., 189 f., 192 f., 199, 207, 209,
amphitheatres,military, 168 f.; in towns, 141, 248, 272, 278, 280, 306, 321, 338, 341 If.,
162, 164,181,227 347 ; seealsofoederati; receptio
annona militaris, 2 I 9 barracks,with dry-stonewalls, 317; wooden, 106
Antiqui barbari, 341 basalt,street-paving,78
apartments,ceremonialof palace, I I I basilicas, Christian, 303, 313, 328, 332, 3341f.,
apron, see costume,local 348,353
apses,in grave-chapels,334 f.; in palace, 3I 3; in baths, in houses and villas, 171, 176, 299 f.;
villas, 171 public, at Aquincum, 161; thermaeLicinianae
aqueducts,159 If.; subterranean,78 at Sirmium, 312
arbiter, II4 beakers,3I 9 f.; motto-beakers,Rhenish, 245
arcade,of houses,164 beams,of houseroofs, 173
Arianism, 329 If. belt-fittings, local, 63
445
Index oj Subjects
beneftciarii, altar of, 254; stationsof, 234 cemetery-chapels,333 fo
berm, of forts, 284 centuriae,tribes organizedin, 70
bishops, Christian, 3 I3, 323, 325 ffo, 341 fo, 345, centuriatio, 78 fo
347 ffo, 351 ffo; and see sees centurions, 242; in control of civitates, 69; as
boars,on pots, 177 praepositigentis, 191
boats,usedin burials, 147 chairs, folding, 148
border-station,260 cheese,producedby Dardanians,26; tradein, 246,
borders, provincial, 33 fo, 93 fo, 96 fo, 137, 198, 299, 322
209, 273 chora, of Aquae, 133,224,303
boundaries,of fields, 7~; and see borders Christianity, Christians, 258 fo, 266, 308, 322 ffo,
boundary-stone,from Cerevic, 114 33 1 ffo, 345, 353
bowls, moulds for, 172; sigillata, 122 churches,seecult-buildings
bracelets,bone, 319 citizens, Roman, from Pannoniancolonies, I I 2,
bricks, clay, 106 II4; in auxilia, 157; new, 138,214;in towns,
bridges, IIO, 2I4fo, 313; building of, 43 fo, 98, 74, 77 fo, II9, 126, 137, 143, 214, 216;
II8, 280 ffo; pontoon, IIO, 287, 295 veterans, 158 fo; see also cives; conventus;
brigands,in Julian Alps, 32 veterani
bronze, 28, 71, 148, 176, 246, 319; badge, 148; citizenship, grants of, 70, 154, 157, 169; pre-
coins, 208; enamelled, 245; gilded, 337; Flavian, 57 fo, 76, 79; Flavian-Trajanic, 100,
mounts,245, 320, 337; statue,313; statuettes, 114, 134 ffo, 141, 151; Hadrianic, 141 ffo, 151;
245; vessels,7, 72, 79,175,245 Antonine, 141; Caracalla,221
brooches, bow-tie shaped 63; cross-bow, 320, cives AgrippinensesTransalpini, 125
337 fo; enamelled,bronze, 245; knee-, 246; Romani,at Brigetio, 126; at Margum, 131
unknown in Moesia, 63; wing, 63, 338 Romaniconsistentes(ad legionem), 126, 139 ffo
buckles, 246, 320 Civitas Amantinorum, 135
building, decline in, 264; techniques,169, 3I 7 Azaliorum, 66
buildings, barrack-likeat palace,3°2;official, 166, Bacensis,166
169; public, II3, 161 ffo, 168; Roman in Boiorum, 66,135,141,171,173
Slovakiaand Moravia, 89 ffo; subterranean in Breucorum,66, 222
cemeteries,3 I 3 Colapianorum,66, 69, 135
burgi (watch-towers),building of, 48, 107, 196 fo, Cornacatium,66, 143
291 ffo, 322 Dardanorum,69, 138
burial-chambers,seeburial-vaults;grave-chambers Eraviscorum,141 ffo, 148, 151, 157
burial-rites, local, 26, 147 ffo, 247, 319; wagon- Hercuniatium,151,222,225
burials, 124, 147 fo, 150 ffo, 159, 175 fo, 234, Iasorum, 143
247 fo Latobicorum, 136
burial-vaults, 238, 315 Scordiscorum,143
businesses,at Aquincum, 161 Varcianorum, 136
civitates,foederatae,57, 70
camilli, on tombstones,248 late Roman, 221, 329
canabae, 139, 142, 168 fo, 226 fo; municipalization liberae, 57, 70
of, 218 ffo, 227, 237; recruitmentfrom, 157 fo; peregrinae, administration of, 49, 69, 134 ffo,
at Viminacium, 168,218,237 I49ffo, 221; contrasts between, 151,222;
canals,constructionof, 109 fo, 272 Dacian, 66; establishmentof, 66 ffo, 134 ffo,
cap, see costume,local 153; groups attachedto, 58; lists of, 53, 66,
Capitolia, of towns, 78, 18 I, 324 68; recruitmentfrom, 5I, 154 fo; relationship
cart-burials,see wagon burials with canabae, 126; relationship with cities,
caskets,mounts on, 320, 337 II4 fo, 135 fo, 141, 143 ffo, 223; territories of,
castra, imperial seatsfurnishedas, 306 49,66,135fo, 138, 141 ffo, 157
cattle, god of, 252; trade in, 28, 31, 126, 299, stipendiariae,57
321 fo classiarii, deductio of, I I 3
ceilings, decorationof, 237 classis Flavia Pannonica,109, I I 3; see also fleets
Cella Memoriae, 348 clay, building in, 159, 166, 175; for pottery, 320
Septichora,335 clientela (client-system),184, 198 fo, 286 ffo
Trichora, 333 fo Codex Theodosianus,221
Celticization, 7, Ioffo, I4fo, 18, 26fo, 55,61 Cohors Batavorum,139
cemeteries,70 ffo, 79, 120, 122, 124, 238 fo; at I Cretum, 96
fortifications, 303, 3°5; at fortresses,70, 77, I Aurelia Dardanorum,195
129, 139; Iazygan,95; late Romanand Chris- II Aurelia Dardanorum,195, 225, 241
tian, 113 fo, 310, 312 ffo, 315, 319, 327,332 ffo, I milliaria Hemesenorumequitata coRo, 195,
348; at towns, 70, 74, 79, II6, 162, 310, 3I 5 227 fo
446
Index oj Subjects
I Noricorum equitata,8I copper,mmmg, 133; trade, 246
I Ulpia Pannoniorum,216 corn, trade, 299, 321 f., 342
I Raetorum,8I corrector, of Savia, 273
I ThracumSyriaca,81, 97, 195 corridors, in houses,164, 171, 173
XXXII Voluntariorum c.R., 81 costume,late Roman, 338; local, 58,63,147,150,
CohortesAureliae Dardanorum,195 152 f., 176, 261, 319
Aureliae novae, 195,216 country-houses,see villas
Latobicorum, 39 courtyards,in buildings, 91, 161, 164 ff., 171 ff.,
Maurorum, 195 300; of gates,303
Pannoniorum,39 craftsmen,71, 122, 124,126, 162, 178,181; seealso
Varcianorum, 39 potters; stonemasons
cohorts, 42, 52, 155; on frontier, 8I, 120; in cremation,238; vessels,17
Moesia Superior, 91, 97 crypt, of cathedral,33 5
coin-dies,moulds for, 28 cult-buildings, Christian, 3I 3, 333, 33 5, 353
-hoards,56, 101 ff., 194, 198,202f., 205 f., 264, cults, see religion
279, 294, 321, 345 culture, native, 147 ff.; abandonmentof, 234, 247,
coins, coinage,Celtic, 28 ff. ; from Apollonia and 25°,259,319,358
Dyrrhachium, 30; denarii, 30, 33, 56, 104, curatorescivium Romanorum,126, 131, 142, 216
206; from easternmints, 321 ; of Eravisci, 56; curiales, in mines-administration,224, 308 f.; of
legionary, of Gallienus, 208; mine-, 131 ff.; towns, 310 f., 317, 323
minted in Pannonia, 208; offering of to Customs (portorium), 200; officials of, 181, 228,
water, 324; the Philippus, 28; rarity in late 231,255 f.
fourth century, 345; of Regalianusand Dry- custor cymiteri, 333
antilla, 206; representingfort-gates,269, 285 ;
Roman, 30,45,49,103,107,294,322; tetra- Dacianization,19, 57,60, 199, 248
drachms,28, 30; see also money damnatio memoriae,of Domitian, !I6
collegia, 125 f., 141, 164, 181 deacons,Christian, 327
Collegium centonariorum,161 decoration, on bone bracelets, 3I 9; on pottery,
fabrum et centonariorum,125 176 f.
funeraticium, 125 decurions, 159, 175; of canabae, 142, 162, 168,
negotiantium,125 2I 8; of cives Romani, 126; of towns, 74, 77,
coloni, 243; leasemines, 134 !I6 ff., 136, 141 ff., 158, 162, 175,214, 221 ff.,
argentariorum, shrine erectedby, 134 225 ff., 237 ff., 244, 249, 263, 308
Colonia (Aelia) Mursa, 222; see also Index 2, deductio, see colonies,foundation of
Mursa deductus(deducticius),!I6 f.
Claudia Savaria, 151, 308; see also Index 2, denarii, seecoins
Savaria diplomas, military, 39, 51,60, 66, 68, 81, 85, 91,
Flavia Felix Domitiana, 136; see also Index 2, 104, 1I2, 142, 155, 157, 175
Scupi dishes, 19, 176 f., 277, 300, 337
Julia Emona,40; see also Index 2, Emona ditches,of forts, 49,106,284;of towns, 74, 310; of
SeptimiaSiscia, 2I 9; seealJo Index 2, Siscia villas, 171
Ulpia TraianaDacica (Sarmizegethusa),136 diversorium (inn), 161
Ulpia (Poetovio), 151; see also Index 2, Poet- dogs, in representationsof Silvanus, 252; pots,
ovio 177
colonies, 70 f., 145, 159, 226, 241, 355; colonia Dolichena,see temples
(nova) at Siscia, 244; foundation of, 40, 50, drainage,162,266,272 f., 298
74ff., 98,112ff., !I8 f., 130, 135, 137, 178; drains, at Aquincum, 310
promotion of towns to rank of, 2I 8, 225 f.; dress,see costume
recruitmentfrom, 94, 154, 157 f., 236; terri- dromos,in tumuli, 151
tories of, 76 f., 79, !I4 f., 135, 158 duces,of Moesia Prima, 294; of Noricum Ripense
colonnades,161, 164, 168 and PannoniaPrima, 348; of Valeria, 291 f.,
columns,depictedon tombstones,186; of Marcus 294, 339
Aurelius, 175, 188; of Trajan, 107 duoviri, of Aquincum, 142, 182
commanders,supreme,of Danubianforces,204 ff. dux et praeses,of Valeria, 340
Concilium Provinciae, 175, 214 dykes, 1I4
Conferenceof emperorsat Carnuntum,268, 276,
302, 308, 324 eagles,on tombstones,125
Constitutio Antoniniana,221 earthquake,at Scupi, 356
consulares,as provincial governors,273, 276 earthworks,271 f., 280 ff.; nearforts, 106
Consularia ConstantinopoJitana,269 economicpolicy, in late third century,265 ff., 272,
Conventuscivium Romanorum,at Margum, 216 297 f.
447
Index of Suijects
embankments,at forts, 107 foundations,of houses,317; of woodenbarracks,
emigration, from Pannoniain fifth century, 327, 106
347 f., 352 ff., 356; from towns in fourth freedmen,acting as vilicus, 124; buried in Aqui-
century, 310 f. leia, 78; of Celtic population, 59, 135; of
'empire of Tylis', 26 Furii, 241, 263; inscriptions of at Margum,
entrep6ts,28, 3I f., 76, 100, 120 216; in mines administration, 131, 134; of
epitaph,Latin, 147, 150; to an Amantinus, 70; to north Italian families, 71, 77 f., 136; in
citizens of Savaria,77; to horse, 139 settlementsat fortresses,141; of Syrian, 227;
EquestrianOrder, decurions, 144, 245; officers, Thracian, 125; of veterans,73, 77
49, 69, 114 f. frescoes,in grave-chambers,3I 3
equitessingulares,143, 222 frontier (limes), end of, 346; in Julio-Claudian
estates,32, 144, 252, 298; Christianity in, 33 5 f.; period, 3Iff., 70; in late Roman period,
development of, 1I5, 173, 235 f., 238 ff., 266 ff., 297, 300, 3I 5 ff., 322, 326, 342 f.; in
299 ff., 307 f., 311 , 3 15, 317, 32 2, 338, 35 6 ; late second-third centuries, 185 ff., 189 f.,
aroundfortified palaces,306; of governorof 194 ff., 201 ff., 208, 211 f., 225, 227, 230 f.,
Lugdunensis III, 342; imperial, 244, 300, 238, 263 f., 307; from Vespasianto Marcus
302 f., 306 f., 317, 356; of Jerome'sfather, Aurelius, 80 ff., II2 f., 129, 157, 176
344; of villas, 124, 171 ff.; see al,-o latifundia fruit, in representationsof Silvanus, 252
evacuation,of Dacia, 209, 2II f., 267 f., 280, 311,
354 game,abundanceof in Pannonia,253
expeditioSuebicaet Sarmatica,102 games,in amphitheatres,169
gardens,representationsof, 335; of villas, 171
garments(Norica vestis),exportedfrom Noricum,
factories, 337; Rhenish, 320 322
farm buildings, at villas, 171 garrison-posts,building of, 43, 80
fauces,in villas, 171 gates,of fortified palaces,303; of forts, 269, 282,
ferry, on Danube,110 284f.; of towns, 74,114,161,310;of villas,
fleets, Danube,45, 8I, 290; of Misenum, 52, I I 3 ; 171
of Ravenna, 52, II 2 f.; see also classis Flavia Genii loci, worship of, 25 I
Pannonica glass-paste,imported, 246
flooding, on Danube,47, 105 f. glassware,found in graves,71, 79; imported, 72,
floors, mosaic, 237, 243, 302 f., 313, 335 320; local production, 177, 245 f.
foederati, barbarian, 296, 339, 341 ff., 349 ff., glaze, on pottery, 177 f.
35 6 gold, brooches,33 7; jewellery, 246; mines, 187;
foedus,with Goths, 339 pin, 353
forests, on estates,303 governors, provincial, 103, 139, 186, 218, 231,
fortifications (fortified palaces), 302 ff., 353; see 245, 264; of Dacia, 99 f., 183 f., 189, 191,
also forts 197; of Dacia Mediterranea,276; of Dacia
fortlets, on Danube,48, 107 Ripensis, 276; of Dardania, 276; of Illy-
fortresses,23, 56,73,85f., 152,212;aqueductfor, ricum, 35; of Lugdunensis III, 342; of
(Aquincum), 162; auxilia stationednear, 50, Macedonia,37,43; of Moesia, 33,41 f., 68,
69; building, 42, 80, 86 ff., 92, 284, 317; ceme- 80, 82; (joint) of Moesia Achaeaand Mace-
teries, 70 f.; centresof trade, 71,121 f., 129; donia, 44; of Moesia Inferior, 82; of Moesia
civil settlements,71, 73 f., II6, 119, 126 ff., Prima, 276; of Moesia Superior,82, 188, 215;
137,139ff., 168,218ff., 223, 226; double, 83, of Noricum, 183 f.; of Pannonia, 37, 41,
85; forts opposite(on left bank of Danube), 80 ff., 183 f.; of PannoniaInferior, 92, 94 f.,
110 f.; foundation of colonies on sites of, II I, 197 ff.; (joint) of PannoniaInferior and
II8 f.; garrisons,48, 69, 82 f., 86, 91 f., 212; PannoniaSuperior, 102; of PannoniaPrima,
of Iapodes,22; see also canabae 273; of PannoniaSecunda,273; of Pannonia
forts, auxiliary, 70, II3, 120, 122, 135, 157, 178, Superior, 94, 103, 186, 188, 199 f., 201; of
201, 241, 279; beyond frontier, 89 ff., 269, Savia, 273
293 f.; civil settlements,143, 173, 175, 194 f., graffiti, Greek, 260
225, 227 f., 231, 238, 247, 252, 260, 315 ff.; grave-chambers,237 f., 305, 3I 3, 3I 5, 334 f.
in Flavian period, 5I, 80 f., 86, 88 f.; hill-, chapels,313, 334 f., 348
native, 5I; in Julio-Claudian period, 48 ff., goods, 58, 315, 319 f., 335
50 f., 69, 73; in late Romanperiod, 269, 277, tablets,Christian, 335
280 ff., 291 ff., 305 f., 317, 346, 349; on left graves,early Roman, 57; of Hunnic princes, 349,
bank of Danube,II 0 f., 337; on Rhine, 120; 352;late Roman-Christian,310,3I 5,319,321,
medieval, 100; in second-third centuries, 325; of martyrs, 348; native, 71, 175; for
88 f., 96 f., 101, 104 ff., 110 f., 194 f. cart- and wagon-graves,see burial-rites
forum, of colony, 314 gravestones,see tombstones
448
Index oj Subjects
gromatici, Roman, 78 jet, carvingsin, 245; import of, 320
ground-plans,of Christianbuildings, 332, 335 ; of jetties, floating, IIO
housesand villas, 171, 173, 175, 237, 300 jewellery, gold, 246; local, 58, 63; silver, 175;
guard-posts,on Danube, 36; on road, 5I trade in, 28, 246
gymnastics,175 jugs,319 fo
hall, 313 kalendariumSeptimianum,219, 244

harbour,of Sirmium, 114 kiln, pottery, 176
hares,on pots, 177 knives, in representationsof Silvanus, 252; vine-
harness,ceremonial,148 dressers',299
hasta (lance-shapedbadge), 148
hearths,of houses,173, 310 lamps, earthenware,122
helmets,gilt, 337 landing-places,on Danube, IIO, 269, 291, 322
herbs, trade, 246 land-tax, in PannoniaSecunda,348
hippodrome,at SremskaMitrovica, 3I 2 language,Greek, 260; Latin, II7, 199, 201, 217,
hoards, at Aszar, 175; fourth-century, 277, 300, 260 fo, 263, 338, 354 ffo, 358; lingua Bessica,65,
3I 9, 32I, 337, 342; seealso coin-hoards 35 8 ; local, 4ffo, 12, 27, 59 fo, 65, 259 ffo, 358
hood, see costume,local lararia, 252
horrea, at fortifications, 303, 305; of mines- lard, trade in, 299, 322
administration, 134; in towns, 308, 312; at latifundia, 138, 238, 241, 305 fo, 315, 319, 335
villas, 171, 302 latrocinium, 234
horses,on tombstones,147 fo; trade in, 246 latrones, 78,195; latrones Dardaniae, 153, 195
horticulture, god of, 252 latrunculi, 196
hospital, of fort, 195 law, instruction in, 262
houses, 319; of Roman troop-commandersbe- lead, mining, 133, 217; reliefs, 254; trade, 246
yond frontier, 89 ffo; in towns and settle- leather,workshops,161
ments, II6, 161 ffo; 237ffo, 299ffo, 310, 312, lector, of Christiansat Cibalae, 325, 327
317; non-urban, 171 ffo, 224, 310; see also legates,see governors
villas legatuspro praetore,in commandof Moesianforces,
house-urns,Latobican, 63 44; see also governors
hunting, -castle, imperial, 303, 306; dogs, 246; Legio I Adiutrix, 82, 86, 92, 99, 118, 198, 243
god of, 252; sceneson mosaics,303 II Adiutrix, 82, 86, 92, 99,118,141,155,168,
huts, at Halimba, 175; of Dardania, 26; at 183,195,198,201,228,243
Savaria,78 V Alaudae, 81 fo, II3, II6
hypocausts,IIO, 164, 176, 239, 310, 3I 3 XV Apollinaris, 40, 43, 48 fo, 69, 74, 77, 80, 89,
92, 94, 99, 155
illiteracy, in Severanperiod, 262 fo III Augusta, 194
IlIyricization, of Scordisci,66 VIII Augusta,43, 48, 74
immigration, 178, 247, 261; Christian, 325; VII Claudia,48,51,65,81fo, 91, 96, 99, II6 ffo,
oriental, 195,227ffo, 238, 247, 255, 258,322, 155 ffo, 188, 243, 249
325 XI Claudia, 92
imperial cult, provincial, 175, 214 IIlI Flavia, 82, 86, 91 fo, 96, 99, 101, II8, 183,
impluvia, in villas, 171 188,198
inco/ae, 126, 154, 158 VII Gemina,48
industry, 79, 176 ffo, 242 fo, 245 ffo, 264, 319, X Gemina, 48, 92, 99, 124, 162, 200, 243
357 XIII Gemina, 48, 77 fo, 88, 92, 96, 209, 212
inhumation, 26, 238 XlIII Gemina, 85, 92, 94, 99, 243
inn, see diversorium IX Hispana,43 fo, 48
insignia, triumphal, 34, 41 fo; and see badge I Italica, 83, II6
insulae, in towns, 74, 161 II Italica, 184
intaglios, depictingVictoria, 72; on rings, 320 IIn Macedonica,II 6
interpres officii coso salariarius legionis, 199 V Macedonica,43, 47, 69, 83, II 6, 187,209,212
iron, Norican, 3I fo II Parthica,200
islands,in Danube,191 XXII Primigenia,48
iter Scorfularum (Scrofularum), 47; see also XXI Rapax, 84 fo, 95
Scrofulae IIII Scythica, 43 fo, 47 fo, 51, 69
iudex,II4 XXX Ulpia Victrix, 92, 94, 99
ius commercii,of Marcomanni,71 LegionesIllyricianae, 208
Italicum, grantedto towns (?), 218 legions, connectionwith towns, 239 fo; Dacian,
iuventuscolens Iovem Dolichenum,181 95 f., 259, 3I I ; defeatedby Sarmatians,294;
ivory, sculpturesof, 245 in Julio-Claudianperiod, 36,41 ffo, 73; under
449
Index oj Suijeets
Domitian and Trajan, 85 ff., 94 ff., II8, 120, mounts,bronze, 148, 246, 320, 337
124, 129 f., 157;in post-Trajanicperiod,98 f. ; municipia, excavationsin, 159 ff.; foundation of,
in late secondandthird centuries,194, 197 ff.; 74,115,117f., 133, 135 ff., 139 ff., 196,2.13ff.,
in Orient, 228; influence on religion, 181; 2.18 ff., 23 I; promotionof to rank of colony,
pseudocomitatensian, 314 f.; recruitment,94, II9, 218, 22.5 f.; recruitment from, 155 ff.;
II7 f., 154 ff., 183 f., 236, 239 ff., 248, 307, Romanization, 151 f .. 231, 355; territories
346 ; veterans,77, 82, 92, II3, II6 ff., 141, of, 78, 135 f., 140, 142 ff., 147, 151, 153,
162, 239 2.14 f., 221 f., 2.24
lime, 109 Municipium Aelianum, 133,2.24
limes, seefrontier Aelium Aquincum, 141; see also Index 2,
limestone,tablets, 246; tombstone,175 Aquincum
linen, found in graves, 321 Aelium Carnuntum, 141; see also Index 2,
liturgy, Thracian, 358 Carnuntum
Iituus (augur'sstaff), 325 Aurelianum, 224, 3°°;see also Index 2, Aure-
loculi, 33 5 liana
locus orationis, seeoratory Aurelium Augustum, 2.16; see also Index 2,
Margum
macellum(food market), 161 Brigetio Antoninianum, 221; see also Index 2,
magical texts, see texts Brigetio
magistermilitum, Aequitius as, 292. Celegerorum,22.3, 225, 23 I
utriusquemilitiae, Maiorianusappointedas, 340 Dardanorum(Dardanicum),133, 223 ff., 276
magistrates,of canabae, 142, 168, 218; of cives Faustinianum,144,222,225
Romani, 126, 142; of towns, 2.21, 231, 308, Iasorum, 136, 143 f., 152, 2.23
32.4 Latobicorum, 135 ff., 143, 151, 223; see also
manufacturers,foreign, 12.4, 129, 131 Index 2, Neviodunum
marble, altar-slabs,336; encasingwalls of build- Mogentiana,144; seealso Index 2, Mogentiana
ings, 302., 313; reliefs, 2.54; sculptures,303 Mursella, 144; seealso Index 2, Mursella
market-hall, 3I 2.; and see macellum Scarbantia,151; seealso Index 2, Scarbantia
-place, at Carnuntum,126, 168 Ulpianum, 133, 138, 143, 145, 153, 2.2.4; seealso
markets,Romanon Danube,191, 193, 197 Index 2, Ulpianum
merchants,seetraders Volg. (...), 2.22, 2.2.5
metalla Aeliana Pincensia,131, 133, 216, 2.23 f. murals, III, 124, 335
Aureliana, 131, 133, 2.16, 224 musical instruments,of Dardanians,26
Dardanica (metalla Municipii), 131, 133 f., 216,
2.23 f. nave, of crypt, 335
Ulpiana, 131, 133 f., 145, 2.16, 2.2.4 navigation,of Danube,45 ff.
metallarii, 195 navy, recruitment,52
milestones,107, 264 neck-charm,63
mines,mining, Dadan,187; in Celtic areas,3If.; negotiator Daciscus,130
Upper Moesian, 63, 131 ff., 138, 153, 188, negotiatores,see traders
195, 208, 2.13, 2.16 f., 2.40, 307 ff.; see also Norica vestis,see garments
metalla; territoria metalli
mints, 321, 343; Sisda,208, 343; Viminacium, 208 ojjicinae, of mint, 343
missio honesta,of Ravennafleet, I I 3 officium consularis, 199
missionesagrariae, I 17 f. oil, trade, 2. 8
nummariae,77, II8, 126 -presses,302
Mithraea, 181, 209, 255 ff., 324 oppida, native, 19, 28, 72 f.; on Mount Gellert,
monasteries,Oriental, 358 72 f., 86, IIO, 173
money,circulation, 30, 1°3,2.82.,321,343; depre- OppidumScarbantiaIulia, 74, 76, 78, I 15; seealso
ciation of, 104; seealso coins Index 2, Scarbantia
monuments,Christian, 312. f., 32.9, 332.; funerary, oratory (locus orationis), 328
2.37 f., 260, 336; late Roman, 337; of legio orchards,at villas, 171
IIII Flavia, 99; native, 153; official, 2. 3 I ; ordo, in third century, 244; of canabae,218; of cives
religious, 181, 250, 254, 2.56, 258; votive, Romani, 126; of colonies, II4, 226; or muni-
26o; seealso tombstones cipia, 141 f., 144, 151, 153, 214, 218, 221;
MonumentumAncyranum,33, 35 Syriansin, 228 f.; veteransin, 239
mortar, 169, 317 organ, of collegium (Aquincum), 161
mosaics,III, 244, 2.99, 302 f., 315, 334, 337; and ornamentatriumphalia, 103
seefloors ovatio, of Domitian over Sarmatians,85
moulds,for bowls, 177 f.; for bronzeobjects,246; ovens,in houses,173, 310
for coin-dies, 28 oxen, on tombstones,147
450
Index oj Suijects
paganism,in late Romanperiod, 290, 323 if., 332 procurators,of mines, 131, 133, 308; provincial,
paintings,in tombs, 3I 5 ;seealso wall-paintings 41
palaces,imperial, 202, 295, 300 if., 312 if., 332; of protectores,Illyriciani as, 2 I °
governor, III, 237; see also villas publicumportorium Illyrici, 200, 255
palaestraequipment,found in graves, 148, 150, punches,177
175 f.
paving, of streets,78 quattuorvirate,of municipium, 144
peregrini, 58 f., 61, 63, 79,137,148,217
peristyles,of market, 161; in villas, 300 ramparts,at forts, 91, 106 f., 284
Philippus, the, see coins reading,teachingof, 262 f.
piers, of aqueduct,159; of bridge, 215 receptio, of barbariansinto empire, 186, 189, 192,
pig's head,on gravestone,261 2°9,268f., 278 ff., 288, 290 f., 296, 339, 342,
pilasters,on tombstones,180 345
piles, of bridge, IIO recruitment, see auxiliary troops; legions; prae-
pin, gold, 353 torian guard
pitcher, on tombstones,148 RegnumNoricum, 76
pit-dwellings, 175, 317 Thraciae, 68
plague, 187, 210 relics, of martyrs, 327 f., 348, 353
plaque,dedication,74 reliefs, bronze, 148, 337; on column of Marcus,
plates,on tombstones,148, 261 175,188;on columnof Trajan, 1°7;cult, 25 3,
plinth, of statue,139 33 6; in Mithraeum, 256; of Rider-gods,254,
porphyry, quarry, 326; statues,303, 337 324; of Silvanus, 250, 252; and see tomb-
portico, IIO, 126 stones
portorium, see customs religion, cults, 142, 153,181 f., 232 ff., 25off., 308,
portraits, gilt-bronze, 337; on tombstones,18o 322 ff., 325, 336, 338; see also Christianity;
posterns,of Emona,74 paganism
post-holes,of forts, 106 res privata, administrator,306
pots, on tombstones,18o rings, 320
potters, 176 ff.; 'first master', 177; see also crafts- roads,amberroad, 71 f., 74, 76 f., I 15, II7 f., 120,
men 122, 152; army units stationed on, 50 f.;
pottery,Aco-beakers,72; barbarianproductionof, building, 43 ff., 100 f., 109, 280 ff.; Danube
247; Dacian, 19; found at late Romanforts, road, 45 ff., 96 ff., 105 ff., 109, II9, 153,225,
282; industry, 176 ff., 245 ff., 319 f.; Iron 284; developmentof towns and settlements
Age, 176; -quarter at Aquincum, 162; red along, 72, 74, 98, II5, 117 ff., 124, 137,143 f.,
andwhite, 72; Roman,71, 176; samian,49 f., 196,213ff., 225, 241; tradealong, 71, 120 ff.,
86,114,123,177,193,245 f., 32o;sigillata, 72, 137,152,24 6
79, 88, III, 120, 122, 129, 161, 177 f., 320; road-stations(praetoria), 100, 106, 215, 298
workshops,161, 178 robbers,seelatrones
praefecti,of alae, 50, 114 f., 135; auxiliary, 69, 189; Roman buildings beyondfrontier, 89 ff.
of cohorts,69; legionary,86, 264; military of Romanization,70 f., 79, II3, II7, 124, 130,147 ff.,
civitates,69, 134 f.; nativeof civitates,134 f.; 176,181 f., 234 ff., 247 ff., 253, 256, 259,263,
praetorian,50, 82 f., 188, 191, 197,210,294f., 31 9 f ., 336, 338
310; ri pac Danuvii et civitatium duarum Boiorum roofs, of houses,173
et Azaliorum, 69; vehiculorum, 101 rosettes,on tombstones,180
praefecturae,military of civitates,69 f., 134 f.
praepositi, native in Dalmatia, 135 sailors, settlementof, I I 3f
praesides,273, 276, 310 SanctiQuattuor Coronati, 326 f.
praesidia, on Danubebank, 196 f. sanctuaries,seetemples
PraetenturaItaliae et Alpium, 187 sarcophagi,237 ff., 246, 264,310,325,334
praetoria, see road-stations schools,262
Praetorian guard, praetorians, 117, 214, 222; scribae, municipal, 144 f., 153, 222, 263
Cotini in, 248; Illyriciani in, 200 f., 204 Scrofulae, rapids on Danube, 45; see also iter
prata legionis, 140 Scorfularum
priests,Christian,326 f., 332; of Dolichenus,256, sculptors,foreign, 181, 337
32 5 sculpture,2°9,244,336;ivory, 245; marble, 3°3;
pr(inceps) c(ivitatis) B(oiorum), 135 porphyry, 303; stone, 100, 124 f., I79ff.,
praef(ectus)Scord(iscorum),134 f. 234, 246, 297, 3 I I
principes,native, 7°,134ff., 141 f., 159, 175 sea,productsof, 28
primipili, 69 sees,of bishops, 214, 222, 303, 310, 312, 328 if.,
primipilus leg. V. Mac. praefectuscivitatium Moesiae 336,35 2
et Treballiae, 69 serfs, of Dardanians,27
451
Index oj Suijects
settlements,civil, seeforts, fortresses;late Roman tabularii, of civitates, 141 f.
fortified, seefortifications taxation policy of Probus,z95, 310
SevenVirgins, 327 temples(shrines,sanctuaries),Z5 3; in Aquincum,
sewers,sewagesystem,78, I z6 161 f.; of Antinous,134; building, 181, z31 f.,
shacks,in canabae,168 Z37; of Dolichenus, Z56 ff.; of Isis, Z44;
she-wolf, on gravestones,151, z6z pagan, 3Z4; of Silvanus, Z 5 If.; templum
shipping,godsof, IZ5, 136; on Danube,109, IZZ; Provinciae, 175; see also Mithraea
on Save, II3, II5, 130, 136 f. tendrils, on pots, 177; on tombstones,180
shop-keepers,Iz6 terracottas,177 f.
shops,at Aquincum, 161; at Sirmium, 3 I Z terre pisee,walls of, 175
shrines,see temples territoria metalli, 134, 195, zZ4 f.
silk, found in graves,3ZI tetradrachms,see coins
silvae dominicae, 306 textiles, found in graves, 32 I ; trade, z46
silver, export, z46; hoards,300, 337; mining, 3I, texts, magical, 324
133, ZI7; works in, 175, z45, Z77, 300, 337 theatres,in Savaria, 328; in Scupi, II6, 181; in
skins, trade in, z8, 3 I Sirmium, z94, 310
slaves,58; of Boian aristocracy,151; of customs, thermae,see baths
Z55; fugitives from Radagaisusseized as, Thracianization,of Scordisci,66
348; ofItalianfamilies,z8, IZI, 136; inscrip- tiles, in building, 89 ff., 109,169,z9z, 31Z f., 317;
tions of at Margum, Z I6; on latifundium, Z4 I ; found at Szeged,100; graffiti on, z60, z6z;
in mining settlements,134; municipal(vilicus), graveswith, 315 ; local productionof, 176 f.;
Z I9; of native new citizens, 135; Pannonian stamped,88 f., 96, 98, z69, z85, z9I ff.; sun-
youths sold as, 34, 37,55; trade in, z8, 31, dried, 78; with Chi-Rho monogram,336
59, IZ9, 321 f.; of veterans,73, 77 tile-works, military, 9 I
smugglers,196 timber, see wood
springs, 324; thermal, 303 tips, from mines, 133
stable, of inn, 161 toga, worn by decurions,308
staff, augur's,see lituus tomb-altar, 143
staging-posts,169 chapel, Christian, 303
stamps,potters', 176 f.; see also tiles tombs, inscriptions on, Z31; stone-built, 335;
state cameo,from Kusadak, 300, 337 subterranean,3 I3, 335
statues,of Aesculapius,3z6; of Capitoline Triad, tombstones,IZ4, zz8, Z37 ff., z43 f., z64, Z97, 3II,
18 I, 3I 4, 324; of Constantine, 3 I3; cult- 315, 323; at Aquincum, I6z; with astral
statues,336; imperial, 337; inscriptions on, symbols, 61 ff., 147, 179, z34, z48; of auxili-
Z3 I aries, 49 ff., 55, 86; of citizens of colonies,
statuettes,bronze, z45 I I 3; of decurions,74, II6 f., 175, zz3; depict-
stelae, 180 f., z64 ing native costume, 58,63, 147, 150, Ip f.,
stipendium,paid to Roxolani, 100 176; of freedman,78; of legionaries,71, 9Z,
stone,building in, 80 f., 106 ff., I6z, 169, 175, 181, 130, 157; of native principes, 134 f.; of
z37 ff., z8z, z99, 3IZ f., 317, 335,346; paving, natives,61 ff., IZ4, 147 f., 150 ff., 176, 179 ff.,
107; votive, I6z, z66, 324; seealso sculpture; z34ff.; reliefs on, IZZ, 147 f., I50ff., 180,
tombstones zI6, z48, Z6I; of Rhenishimmigrants, IZ5;
stonemasons,86, 319; Dacian, 3 II; decline in of sailor, I 13; of scriba, 144; of traders, 7 I,
standards,z64; foreign, 179; in late Roman 73,78 f., IZO f., 130, IP; of veterans,49, 73,
period, z97; local, 15 z, 179 ff., 336 f.; mar- 77, 8z, II6 ff.; of woman interfecta ab hoste,
tyred, 3z6 f.; in mining area, Z I 6f.; in pre- 188; at Viminacium, lz9 f.
Flavian period, 50; in towns, IZZ, IZ4, 130 tools, for coin-striking, z8; for making sigillata,
storage-containers, in imperial palace, 30Z 177
stratores, of governor of Moesia Superior, Z15 towers, of fortifi.ed palaces,303, 305; of forts, 88,
street-grid (system), in civil settlements,I z6; in 107,194,z69, z80ff., z9z; of villas, 300; see
towns, 74, 78, II6, 159 ff., 164 also burgi
streets,in canabae,168; in towns, 78 f., 113, 161 ff. trade, acrossfrontier, Iz6 ff., 193, 197, z43, z45,
strigils, on tombstones,148 z47, z93, 3zz; early Roman basedon Aqui-
stucco, III, 177, z37, Z99 leia, 3 I f.; export, I z9, 178, Z46 f., Z99, 3 Z If.,
symbols,astral on tombstones,61, 147, 179, z34, 34Z; in Flavian-Trajanicperiod, 117, 119 ff.,
z47 f. ; Christian on sarcophagus,334 137,15I f., 177;infourth century,Z99, 320 ff.,
synagogue,at Intercisa, zz8 34Z; Greekin Danubearea,z; import, 70 ff.,
79, IZO, IZZ ff., 148, 151, 176 ff., Z37, Z45 ff.,
tablets, votive, Z46; wax, from gold mines, 187 z99, 3zo f., 34Z, 358;in Julio-Claudianperiod,
tablinum, in villas, 171 35, 70 ff., 78 f., in late Celtic period, z8 ff.; in
tabula, Trajanic, 109 late second-thirdcenturies,z43, z45 ff., z64
452
Index oj Suijects
traders(negotiatores),126, 130, 137, 152, 182, 216, Mediana, 286, 291, 302, 324, 337; Orvenyes,
357; Italian, 30 f., 33, 71, 73, 76 ff., II3, 336; Parndorf(=Murocincta?),135, 171 ff.,
II9 ff., 126, 130, 136, 173; Syrian, 141,229 238,244,295,299,302; Remesiana,300, 335;
trading-posts,betweenHuns and East Rome, 350 Romuliana, 302; Smarje, 171; Winden am
transport,see water transport See, 164, 173
trees,in representationsof Silvanus, 252 vine-leaves,on pots, 177
tribunusgentis Marcomannorum,345 vines, cultivation of, 266, 298 ff.
tribus, mentionedin inscriptions, 219 vineyards,at Aquileia, 32
Claudia, 77
Velina,77 wagon-burials,see burial-rites
tripods, on tombstones,148, 248 wall-paintings, 237 ff., 299
triumph, of Domitian over DaciansandChatti, 84 walls, of amphitheatre,162; of canabae,169; dry-
tumuli, 71, 79, 148 ff., 234, 247 f. stone, 3 I7; of forts, 106 f., 110, 269, 282 ff.;
turban-veil, see costume,local graffiti on, 260; of grave-chapel, 335; of
twins, on tombstones,15 I, 262 houses, 169, 173 f., 237, 239, 317; of late
tympanum,on tombstones,180 Romanfortified settlements,303 ff.; of palace,
3 I3; of Roman building in the Barbaricum,
utensils,sacrificial, on tombstones,148, 150 91; of tombs, 313; of towns, 40, 74, 78,
113 f., 126, 161, 164, 166, 168, 215, 294,
vases,on tombstones,180 3Iof ., 344, 348, 357; of villas, 171, 302
vaults, of grave-chapels,334 wastematerial, from mines, 133
veil, seecostume,local watch-towers,see burgi
vestibulum,171 water supply, of towns and settlements,78, 126,
veterani et civesRomani, 158 162
consistentesad legionemII adiutricem, 126 water transport,35,45,100,109f., 131, 177, 1-46,
veterans,citizenship grants to, 137, 142; settle- 266, 298, 322
ment of, 73 ff., 78 f., 92, II6 ff., 124, 126, wattle and daub, 168
140 f., 144, 158 f., 173, 175, 239; see also weapons,trade in, 28
auxiliary troops; legions; tombstones wells, 126, 324
vicar, of Illyricum, 214 wild animals,trade in, 246
vici, 76, 115, 176 wine, Pannonian,298 f.; trade in, 28, 299, 342
vicus losista, I 14 wine-presses,299, 302
lovista, 143 wood, building in, 106 ff., 159, 162, 166, 173 ff.;
vilici, administeringKalendariumSeptimianum,2 I9; working, 179
of estate-owners,124; of road-station,100 woollen goods,producedby Dardanians,26
villas, 91, 159, 166; alae in, 171, 300; Christian workers, agricultural, 103
buildings at, 325 f.; early, 124, 151, 164, workshops,in canabae,168; central, 277, 302, 320,
169 ff., 176, 239; in fourth century, 299 ff., 337; leather, 161; metal, 148, 161,277,320,
306, 310 f., 337; imperial, 286. 291, 295, 337; mosaic,237, 302, 3 I3; pottery,161,178,
300 ff., 324, 33 7; palace-type,17I, 238, 244, 320; of Roman craftsmen, 7 I; of stone-
300 ff.; in second-third centuries, 238 ff., masons,152, 179 f., 197, 262; in towns, 152,
243 f. 161, 166, 178, 237, 337
Balaca,171, 307; Deutschkreuz,299; Donners- wreaths,on tombstones,180 f.
kirchen, 336; Eisenstadt,171; Kostol, 300; writing, 26 I ff.
453
LCANNABIACA
IKlosUrneuburg
VINDOBONAI
len
Legionar y fortres s * f'\^ ALA NOVA ^Rusovce^Rusovce
rd ><$chwechat I \yOeutsch-Aìtenburg, Petronell
(2 n d - 3 Century )
Fischämend1
Regefsbrunn/ GERVLATA
F o r t o r cam p ^Rusovce Ipo/y
dor f
Winden'cùn)Pam
Coloni a o r Municipiu m .See/, •A
) D onnerskirchenÁ
AD F L E X V I ^ » '
Othe r localit y /Eisen stadt MagyarSvár \ AD MVB.ES C \ Hidegieloskereszt PHismarot-52 ob
^Rusovce
a *+Jfospuszta ADs T A? V A 5"ELAMANTIA
| , / Vìsegraà
" ¡th rMattersdorfS QVADRATAI tèa I AZAyM
Lébény Acs_ I Almasfù'z/ió "{Dunaboódány
Road
SCARBANTIA F O L V A ^
,TKj'.'rSopron ' yARRABONA CRVMERVM *
• ^^Rusovce
Bridge 6RIGETIO IP VLCISIA CASTRA
SSfSySr Szöny ^RusovceFEZterg>m( 757/ / Szentend re • Haivan
" T O - ßudaka/aszm
Summi t Ì) MVRSELL/S • Környe
1 ^Rusovce ÂAVINCV M
Mórichida- mÁszár
\ Gyalóka Kisàrpàs • Budapest
Lan d ove r 30 0 metre s Rtp^ AlbertfaJm
882 r t-CAMPON A
) Nagytétény
ii a 50 0 u 5 AVA RI A Vi I '
\Szo/n6athefy ( 3 ^Rusovce
¡¡/MATRIC A
c
11SzázhaJomb
atta
MOGENTIANA i Pogánytelek ( ì 1i
J TüskevárV-
^Rusovce • G O R S I VM
^Rusovce
8aláca * • Szentkiralys2abadj<ri ilVETVS SAUNA
mB.-Fokajárli \\Adony e
mHa/imóí^
ßrFüredi Polgárdñ vc
SümegcsehL¿ Tihany* INTERCI5AB BDunau¡varos
^Rusovce
e uso
Rezifr 'Kékkut \Sarszentmìk/ós1ì
ì>SAL> T AN N AMATI A v^cR ce
Zalalòvo T R I C C I A N AÌ ßaracs so ov
^Rusovce B.-Györök£ Sfyär u s
^Rusovce ^R u
^RusovcesFenékpuszta ^R
||\\LVSSQNIVM// J
Dunakom/öd^ß
J e
Jill 7/*m
vc
Jj POETOVI& II uso
tìetenypuszta //ALT A RIPA
/ PtojÂ •/O . ToIn a. ^Re
CELEIA ' c
ATRANS_ 'momarje Szekszárd• ov
Szalacska • Il
AQVAE IA5AE m r
us e
e
Var. TopJice JOVIA •682 . • AD^STATVA S ^R vc
ovc
• Komìo, \\ Vardqsnb so
LjubìjanaJ&h^
o
1+643 . Hossiuhetény
v^cReus
u
u s
e R ^R
. P R A E T O R I V M^ vc ^
' N A V P O R T VS ^LATOBICORVM / MVNICIPIVM LATOBICORVM// VGIOB o
^Rusovce
' DunaszekcsóVJt
e
\NEVIODVNVM // SOPIANAEN ' C O N T RA

^Rusovce F L O R E N T I AM
v Vrhnika \ •xZrebnje ' .
c
I Pecs N us
$j\Drnovo,Hr£ko / p II
. A/ad/eìk ^R
veJike Maten& ÂLTINVM «
\JCó/ked/i<
uesov
(Í
J) ANDAVrONIA ^l^G-iSCL NOVA5
•f^Síitarjevo Beremendm Zmaievac
so^vRc
5a/ ! ^
^Ru
MVNICIPIVM IA50RVM F 95 3 M V R S E L LA ,
^Rusovce AQVAE BALISAE Petrijevd
^ ^ ^ Darvvor^(5)
FLSLSCIA
^Rusovce MVRSAft \
Osijek T E V T O B V R G I VM
^Rusovce
^R
us ^CIBALAEJÌ I C O R N A C VM
ov VinkovceG • ySo^/ 7
ce BONON!^"
ScLUct ßanoStor^ß
^Rusovce
^Rusovce ^Rusovce
^Rusovcefcortanovei'SJÌACVMINCVM
^Rusovce ^m^lankamen
B V D A L IA Beáka ^kRITTIV M
S I R M^Rusovce B A S S I A NA
I V M, ^Rusovce
Mitro vica
zQPetrovce • F ß V R Q E N A E
^Rusovce
M F O S S AE \\V/VoK/' Banovci
Jarak.
Hrtkovci
TAVRVNVMJ
Zemun r
0 50 1oo 200
? /"^5INGLDVNV M
KM ' ßeograd
PLATE Ia(above) Greek bronzevesselfrom the Celtic cemeteryat Szob (p. 7)
PLATE lb (below) Late Celtic vessel with Latin inscription DA BIBERE from the
Romano-Celticcemeteryat Cserszegtomaj(p. 124)
PLATE 2 Sometypes of barbariancoins (obverse)(1/1) (pp. 28-30, 56)
PLATE 3 Reversesof coins on Plate z
PLATE 4a (above) Norican-Pannonianfibulae (p. 6,)
PLATE 4b (below) The impressionof an intaglio showing Victoria on a late Celtic
red-paintedvesselfrom Aquincum (4/r) (p. 72)
PLATE ~a (dght) TombstoneofScerviaedus,son of Sita, slain by robbers(pp. 6j, 180)
PLATE jb (centre) Tombstoneof Atta, son of Bataio, a trader from Savada(pp. 78f., ljZ)
PLATE jC (left) Tombstoneof Tiberius Sattaof Cambodunum(Kempten),veteranof
legio X Geminafrom Aquincum (p. 162)
PLATE 6a (right) Tombstoneof Verondacus,son ofVeruicus from Torokbalint
PLATE 6b (centre) Tombstone of CorneliusZo~imus from Viminacium (P.130)
PLATE 6c (left) Tombstoneof CorneliusRufus from Viminadum (pp. I80f., 261)
(left) Tombstoneof Cornelius
PLATE Sa (above) The late Romanfort at B oljetin in the Djerdap (p. 107)
PLATE sb (below) View of the Ojerdap (p. 4 S)
PLATE (pp
9a (above) The rock-cut Roman road in the Djerdap . 4~ff.)
PLATE 9b (below) Trajan'sroad-inscription(the Tabula Train",) in the Djerdap(p. 47)
PLATE l oa (dght) Tombstone of L. Licinius Lepidus from Dozmat
PLATE lob (left) Tombstoneof Reginus son of Troucetissa of Trier. from Aquincum (pp. 114.)
PLATE I Ia (right) Tombstoneof C. Sextilius Senecio,decurionof Scarbantia
PLATE II b (centre) Tombstoneof SemproniusMarcellinusand his family from Savaria(pp. 179f., 261)
PLATE IIC (left) Tombstoneof Aurelius Runnusfrom Intercisa
PLATE I2a
Tombstoneof
Adius Munatius
from Intercisa
PLATE ub (left) Tombstone of Aelius Septimusfrom Bdgetio
PLATE uc (right) Tombstoneof Bozi daughterof Vellasa from Etcsi
PLATE 1 ~a (left) Stoneslab fr om an aedimlaat BOlcske(p. 1.6t)
PLATE t ~b (right) Coins or Trajan referring to the mines (pp. t} J-})
(i) left, RIC 706, lIIefalli Ulpiani Pan (Z/ l)
(ll) right, RIC 704, Dardonin (1./1)
PLATE 14a (above) The mountainsof the me/alia Ulpia/Jo ncar Ulpiana (Gracanica)
(p. IJI)
PLATE 14b (below) The Danubeat Visegrad(p. 293)
PLATE Ija, b Finds
from cart-graves
(p. '4')
PLATE 16a (above) Aquincum municipium: the !H(Jrtl/llIlI (p. 161)
PLATE 16b (below) Aquincum municipium: the 'big house'(pp. J61, 166)
PLATE 17a, b
PLATE
(opposite) Aquincum municipium: the
so~called 'large baths'(p. 161)
18a (above) Carnuntummunidpium: HouseVI (pp. 161-4)
PLATE 18b (below) Carnuntummunidpium: HouseIV
PLATE I9a (above) Carnuntum:the civil amphitheatre(p. 164)
PLATE 19h (below) Carnuntum:the four-way atch (Heidentor)
PLATE 20a (above) Excavationsat the municipium Dardanorum:the bOrreU!11 (p. 223)
PLATE 20b (below) The Romancity wall of Scarbantiaas reconstructedin the Middle
Ages(pp.166,3")
PLATE 2.I The Aszar hoard (p. 175)
(a) (opposite)Jewelry (b) (above) BronzeVessel I (c) (below) BronzeVessel 3
PLATE zz The military diploma from the Aszar hoard(pp. I j j, 17j)
PLATE 23a The Nymphaeumof
the Sanctuaryat Gorsium (p. 175)
PUTE 23h Sarcophagusof dill8

type from Vimlnacium
PLATE 142 (above) Some types of Pannonianstampedpottery (pp. 176f.)
PLATE 14b (below) Samianware from the 'Siscia' pottery, found at Viminacium
(p. '78)
PLATE 25a A mould of
the potter Pacatus from
Aquincum (p. 176)
PLATE 25b The Emperor

Marcus Aurelius: stamp
for a potter'smould from
Aquincum
PLATE z6a (left) Tombstone of Trebia Lucia from Timacum minus (p. 180)
PLATE z6b (tight) Tombstoneof Aelia Clementilla from Ulpiana
PLATE 2.7a (above) Scarbantia:the CapitolineTriad, Juppiter,Juno and Minerva
(pp. 18',2.44.32 4)
PLATE 27b (below) Savaria:relief fr om the Iseum(p. 181)
I'LATE 18a, b (above) Mosaic pavementsfrom Aquincum (p. zn)
PLATE 19a (oppositeabove) The sarcophagus of Pia Cderinafrom Aquincum (p. In)
!'LATE 19h (oppositecentre) The sarcophagus of the scholasticus L. SeptimiusFuscus
from Aquincum (p. z6z)
PLAn 19c (oppositebelow) The sarcophagus of the lllltrpm Da&orllw from Brigetio
(p. '99)
PLATE 30a Statuette of Negro boy
from Aquincum
PLATE Job Burial-vault from

Brestovik. east of Singidunum
(p. '3 8)
PLATE 31a Stone slab wi th relief of the god
Sil vanus from Aquincum (p. 2 j I)
PLATE 3I b Reliefof Dianafrom

BalatonviIagos (pp. 2.S Iff.)
PLATE}l.a Stoneslab with mythological scene: Bellerophonand the Chimaera.
From Imercisa(pp. Z.6If.)
PLATE 3tb Stone slab with mythological scene: Aeneasescapingfrom Troy.
From Intercisa(pp. 161f.)
PLATE 33a (above) Stone slab with mythological scene: Priam and Achilles.
From Aquincum (pp. 261ff.)
PLATE 33b (below) Stoneslab with mythological scene:Achilles and Hector
(pp. ,6,££.)
PLATE 34a (right) Lead tablet showing the DanubianRider-gods(p. 2. S4)
PLATE 34b (left) Altar erectedat Carnuntumto Mithras by the Tetrarchsin p8 (pp. 2.76, 32.4)
PLATE, sa Glass
vas Jialrellim from
Szekszard
PLATE }Sb Silver vesselof Licinius from Esztergom(pp. 277. }oo)

PLATE 36a Late
Roman fort at
Tokod (p. 3'7)
PLATE 3Gb
Valentinianic
hllrgll1 at Visegcid
(p. '9')
PLATE na, b Gamzigrad: late Roman imperial palace (p. ~o~)
PLATE ,8a (above) The early Christianbasilica at Fenekpuszta(pp. 3°3. H6)
PLATE 38b (below) The early Christian basilica at Ulpianum (pp. H2., 334)
PLATE '9a (above) Sirmium: the late Romanbaths (p. 311)
PLATE '9b (below) Sicmium: pact of the late Romanpalace (p. 311.)
PLATE 40a Sopianae
(Pecs): painted tomb
No. , (pp. l' 3, lJ I)
PLATE 40b Sopianae

(Pees): tomb with
painting of SS. Peter
and Paul (pp. 313.
lJ4)
I'LATE 4Ja(above) Sirrnium: fragmentsfragments
of (screen)from a Christianbasilica,
fragments
built into a modernwall
PLATE 41b (below) Paintedtomb in the Christian cemeteryat JagodinMahab
PLATE 4za. b Details of painted figures in the Christian cemetery of Jagodin Mahala
PLATE 4P (right) Paintedfi gure fr om a tombat Sdka
PLATE 43b (above) Mosaic from the Constantinian villa of
Mediana (po 30:')
PLATE 44a (above) The augur'sstaff from Brigetio (p. FS)
PLATE 44b (centre) Bronze plate from late Roman box: Kisarpas(pp. 3zo, 337)
PLATE 44c (below) Bronze plates from late Roman boxes: Sagvar(pp. 3zo, 337)
PLAn 4sa (left) Silver tripod fr om Polgardi (p. H7)
PLAn 4Sb (right) Late Roman inscribed gold pin from Fenekpuszta(p. 3n)
\\ LEDERATA
>MPa/anka
SINGlDVNVMv ^Rusovce
PINCVM
BeograàZ^/ Hi^X Ve!. Gradiate DI^Rusovce
ERMA.
Orìova\
CTRICORNIVM |VIM1NACIVM / ^Rusovce
MVNICJPlVM/
sJ'P DROBETA
•jv) Kostolac / \|VIM1NACIVM rTekiia. '
CtzavaJra J
an%sTa
MVNICJPlVM/
Turnu-Sevcrtn
CTRICORNIV • M A R G VM |VIM1NACIVM
fOraSje\\\ tàosman j * ^ ^PONTES
M
IVM
.tfrestov/k -(g)- MVNICJPlVM/
IVM
omedertvo C V P P AE
Golubac Sa/dum
CIV
*Hajdutka / sJQadovo
AC
• MVNICJPlVM/
AVREV5 MONSvV
AC
MVNICJPlVM/
>7
68
1NA
1N
q ¡/edenica, lì
1N
*• . Séona \ A MVNIC1PIVM Bo/jetinw,
M
M
Ravnal
C|VIV
|VIM
I
. ¿¡sovic fòKaHste • /
|VI
ITAUATA \\ßrza Pa/anka
• ( E G E T A
NA
^Sioinikm 62&\
M1
\D.MHanovac
M
|VI
CIV
1NA
cAGlVAE?
\Prahovo
|VIM
^Rusovce
/fusadak\
|VI
M1 Negotin )
|V N
|VIIM1 ACIV
1182 ' • Bor,
M1N M
NAAC
CIIVM
" ARGENTARES^-
VM
707 ^ rBONONlA
" ^éptina. N lridi a
• A V R E L I A N A «; ^ • .
\ HORREVM MARGI |VI ^^^J^Zaj ec an/.

M1
Cuprija': : NA AQVAE? '
CI
VMNACIV
M Gamzìgrad
RATIAR1A
|VI
Paracin |VIM
1 M1
NA Arfer ^
1560
.. . .. • • • . .
CI
|VIM1NACIVM
Kraìjevo t
15 ARMATA E |VI VM
MVNICIPIVM/ j *i>PRAESIDlVM DASMINI ' ; ITIMACVM MINVS M1
CELEGERORVM? ^ Ravna-^- » •
NA
IVM r^Ä.PRAESJDlVM POMPEI . |VCI IV
NAC M1M
IM1 -Moravo* NA
|VIM1NACIVM |V CI
1186 : ; VM
\Vrnjalka Banja
•yvanjkaL '
"/ «380 *
.NAISSVS
|VIM1NACIVM 1NA
\A// 5 .
946 ^Rusovce
MEDIANA
2017 I
|VIM
IVM
; R E M E S I A N A ^V
) Topn ca.
AC
+ Beta PaJanka*
r. ; I8O8 / . . . J*
VMM1N
CIV
I
I
IVNMA|V
C
M
AC1
Novipazar/
1N|VIM
MVNICJPlVM/j
ce
.r " '
M
-t - . Leskovac %
MVNICJPlVM/
|VI
ov
. oocaniccL | - AQVA E B A S . .. . 1409

us
Kurìumiìjska
art ¿in Grad
^cRe
v\V. Sonja.,
ov
us
Mitrovica
^R
>1638 ^
\ -4 -
|VIM
|VIM
1N
|VAIM
1NA
C1IVN
• PrìltìncL
M
L e g i o n a r y fortres s
A
CIV
> Pec ILVLPI ANVM

CIV |VIM1
( 2 n d - 3 r d century )
ce
•J+Grafanìca
M NAC
|VIM
v
|VIM
so
1N
MLìpljan AC
For t (o r camp )
1NA
ceu
IVM
ov^R
IVM
CIV
Coloni a o r Municipiu m
M
us
• Orahovac
^R
Othe r localit y
Road
(lesi
*Prìzren
Bridg e
ov ce
ce
Kumanovo% i
ov Russov
. STATI 0 « Summi t
ce
LAMYD Jj.
ov
ê u
^R ovc^cRe
1 L a n d ove r 50 0 metre s
us
us ov
^cRe
rvSCVPj . I " » 100 0 »»

^R Rus
7 Skoplje^
us
^
O SO 100
«748 '
KM
Figure 60 Map of Upper Moesia showing placesmentione<.lin the text

(Routledge Revivals) András Mócsy - Pannonia and Upper Moesia - A History of The Middle Danube Provinces of The Roman Empire-Routledge (2014)

Uploaded by

Copyright:

Available Formats

You might also like

(Routledge Revivals) András Mócsy - Pannonia and Upper Moesia - A History of The Middle Danube Provinces of The Roman Empire-Routledge (2014)

Uploaded by

Document Information

Original Title

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

(Routledge Revivals) András Mócsy - Pannonia and Upper Moesia - A History of The Middle Danube Provinces of The Roman Empire-Routledge (2014)

Uploaded by

Copyright:

Available Formats

Routledge Routledge Revivals

Taylor & Francis Group

Pannonia and Upper Moesia

A Library of Congress record exists under LC control number: 75306910

ISBN 13: 978-0-415-74582-6 (hbk)

LONDON AND BOSTON

Preface page xix

Chapter 2 Conquestof the Danuberegion and Augustanfrontier

Chapter 6 Criseson the Danube:the rise of the Illyrican soldiery 183

FRONTISPIECE Late Roman officer's gilt parade-helmet(no. I) from Ber-

At the time when the Greekswere alreadyfamiliar to someextentwith the geo-

land over 1500

land over 7500

land over 7500

. the second centuryB.C.

land over 7500

land over 7500

noniansinvadedIstria. CassiusDio, who recordsboth theseinvasions,saysthat

Figure 7 The distribution of Greek, Romanrepublicanand native coins

The earliestdistribution of the considerablenumberof legionssentby Augustus

Figure 8 The tow-path in the Djerdap (reconstructionby E. Swoboda)

The principalpeoplesandtribes of pre-Romantimesareall mentionedalsoin the

Figure 9 The native peoplesof Pannonia

attested,2cannotbe located);otherswere called after rivers: Colapiani(Colapis,

Figure IO Tombstoneswith astral symbols

Figure I I Languagesand tribal nameswith Latin suffixes

land over 1500 metres

Figure I 2 The native peoplesof Upper Moesia

be used as a basis for assessingthe transitional period. This settlementon the

Figure I} Town plans I

Experienceson the Danubefrontier in the Year of the Four Emperors,and not

Amphith Late Ron1an Fort

Figure I4 The environsof Aquincum

Figure If Romanbuildings in the Barbaricum

Figure I6 The division of Pannoniaunder Trajan

neighbours,while the Sarmatians,always known as the Iazyges,borderedthe

Figure IS Danubeforts in Pannonia

The spreadoj cities

was thereforenot accidental.There is epigraphicevidencethat the spreadof

Figure I9 The distribution of someItalian families in Pannonia

the Savariabranchof the family is providedby their relatives,the Marcii, who

date date date

Figure 20 Italian terra sigillata in Pannonia

Pannoniavia Aquileia, so that at a fairly early datetheir samianbeganto appear

Figure 2I Inscription set up by the burial-club of the cives Agrippinenses

eagleswith outspreadwings (PI. IOb)). Rhenishornamentthen appearson the

Figure 22 Legionary£ortresses:can baea and municipium

Figure 2} Canabaelegionis: Carnuntumand Aquincum

was vf importancein west-easttrade, particularly in trade betweenItaly and

Figure 24 Mines and municipia in Upper Moesia

considerablenumbersof miners from Thraceand Dalmatia into the province,

Canabaeand Hadrianic municipia

Figure 25 Ulpii and Aelii in the territorium of Aquincum and Ulpianum

Burial rites, tombstonesand their social significance

Figure 26 Monumentsto natives in Pannonia

aboutthe beginningof the secondcenturyin the easternpart of the province.I 37

Figure 27 Auxiliary diplomata from Pannonia

The towns and their archaeology