O!

TRE oo
- IMPERIAL FRONTIERS.
SIR MORTIMER WHEELER
This book is a new and concise survey
of Roman adventuring far beyond the
political frontiers of the Roman world.
Fresh discovery has enlarged our know-
ledge of the astonishing spread of things
Roman from Ireland to China, and
some of the new information given here
has not hitherto been readily accessible.
A part of it is derived trom the author's
own explorations and excavations in
India, Pakistan, and Afghanistan.

255. nel
 SOUTH
WHIDGE CATALOGUED

nee S| rchae ohos

(WH ZELER

21796

TO BE
DISPOSEN
BY
AUTHORITY

AIM
House of Commons Librar

54056001346037
il
 Digitized by the Internet Archive
in 2022 with funding from
Kahle/Austin Foundation

https://archive.org/details/romebeyondimperi0000mort
 ROME BEYOND
THE IMPERIAL FRONTIERS
Bronze statuette of Harpocrates from Tanila
Punjab. + (See p. 158)
 ROME
BEYOND THE
IViIPP Relat PRON PEERS

BY

SIR MORTIMER WHEELER

Professor of the Archaeology of the Roman Provinces
in the University of London

LONDON
G. BELL AND SONS, LTD
1954
 Pie rACE

HIS little book first took shape on a hot May morning

in 1945, when an Indian student of mine emerged
excitedly from a deep trench beside the Bay of Bengal
waving a large slice of a red dish in his hand. Removal of
the slimy sea~-mud revealed the dish as a signed work of a
potter whose kilns flourished nearly 2,000 years ago and
5,000 miles away, on the outskirts of Arezzo in Tuscany.
Were drama admissible to the archaeological scene, I should
have been tempted to describe the moment as dramatic. In
that moment the pages of the historians and the geographers
leapt to life; the long, acquisitive arm of imperial Rome
became an actuality.
Elsewhere in the East, the discovery of Roman things, or
their more local simulacra, has outpaced adequate record in
recent years. The reason lies partly in the accident that few
orientalists have been trained in the classical school, and that
the significant Western material therefore, when found, is
not always recognized. The present book does not profess
to fulfil this need in any comprehensive sense. It is not
an. archacological gazetteer, such as is now badly wanted;
but it may at least serve summarily to indicate the scope
of the problem and to invite further attention to those
matters.
For the West, the field has been admirably surveyed, for
some time to come, by Dr. H. J. Eggers in his monumental
work on Roman imports into Free Germany, published
while this book was in preparation. With Dr. Eggers’s
generous permission, I have drawn freely upon his material,
particularly for my maps. To many others I am likewise
greatly indebted for illustrations: especially to Dr. H. C.
Broholm of Copenhagen, to Dr.G. Caputo, lately head of the
Archacological service in Libya, to Monsieur M. Reygasse
v1 ROME BEYOND THE IMPERIAL FRONTIERS

of the Bardo, Algiers, to Mrs. Olwen Brogan and the Society

for the Promotion of Roman Studies, to the Archaeological
Department of the Government of India, to the late Director
of the Baroda Museum and Picture Gallery, and to the
German Archaeological Institute.
R. E. M. WHEELER

Institute of Archaeology,
University of London, 1953
 CONTENTS
Page
Preface : : : ; : : nny.

I. Introduction : t : : : 3 I

PART I “EUROPE

II. Free Germany: The Literary Evidence . : : a

Ill. Free Germany: Routes and Markets. . See

IV. Circumstances of Discovery . ; ; : rgt

(a) Liibsow ‘chieftains’ 32 (b) Other burials 45
(c) Settlements 48 (d) Loot 53 (e) Peat-bog
or moss deposits 54

V. Free Germany: The Imports p03

(a) Coins 63 (b) Vessels of silver and gold 68
(c) Bronze ware 72 ~— (d) Glass 84 ~— (e) Terra
sigillata 87

VI. Free Germany: Summary . ; : : 2 OF

PART I~ APRICA

VII. The Sahara ‘ : : : : : OR

(f) The Fezzan: ; ’ ; ; 07,
(ii) Tin Hinan : , ; F , BLOF

VIII. East Africa . : , : : : : eeert2

Viil ROME BEYOND THE IMPERIAL FRONTIERS

PART Il Asi”
Page
IX. The Periplus ; ; : : : , ets

X. The Monsoon . : , : ' 5 120

XI. From the Indian Standpoint . ‘ : ; kde

XII. South India , ; , ; ; : 237

XIIL Pakistan and Afghanistan. : ; : ey aes

XIV. The Far East . - ; : : : 2 S72

XV. Retrospect . ; ; : : : ; = “176

Select Bibliography . : : é ; nae ee

Index : : : : : ; , ales
 IEPLUS TRATIONS

PIGURESAN THE TEXT

Page
. Map of principal European trade-routes 13
. Map showing differential distribution of Roman bronze
vessels and coins, lower Vistula region . 25
. Map of ‘chieftains’ graves of the Liibsow group 33
. Reconstruction of Early Iron Age homestead at Ezinge,
Holland 51
. Distribution of Augustan gold coinage resulting pos-
sibly from the Varus disaster, a.D. 9 55
. Map of Roman coins in Free ca ee to
AD 2s0 65
. Map of ‘dolphin’ situiae ae swan’s head’ paterae 74
. Map of paterae with circular piercing 76
. Map of ‘the poor man’s bucket’ a7
. Map of ‘Hemmoor’ buckets 79
. Map of ‘Westland’ bronze vessels 8I
. Map of fluted bronze bowls 83
. Map of terra sigillata ‘ : 88
. Plan of the ‘palace’ and tomb of Tin Hinan 108

. Map of India, showing ne and markets in the Ist

century 7aX D5 119
. Map of Roman coins in ee ‘ 138
. Map of Roman coins of the rst aS A.D. in South
India 144
. Rouletted dish from ace nk ade 149
. Roman glass from Begram, Arikamedu and Taxila 159
 PLATES

Frontispiece. Bronze statuette of Harpocrates from Taxila, Punjab.

Photo, Archaeological Survey ofIndia.
Page
i A. Silver cup from a ‘chieftain’s’ grave at
Hoby, Laaland. 20
B. Silvercup, partially gilt froma ‘chieftain’s
grave at Dollerup, Denmark 20

IL. Bronze jug from a ‘chieftain’s’ grave at

Hoby, Laaland 21

Ul. Roman bronze vessels and glass with

drinking horns etc., in a ‘chieftain’s’ grave
at Juellinge, Laaland 36
. Roman vessels of bronze and glass from
‘chieftain’s’ graves at Juellinge, Laaland . 36-7
. A, Silver cup, partially gilt, from the
‘chieftain’s’ grave, LibsowI . 36-7
B. Glass cup with silver overlay Gon
Varpelev, Zealand 36-7
VI. A. Bronze statuette of Mars, from
Tybjerggaard, Denmark 37
B. Bronze statuette of ee from Ezinge,
Holland a7
VIL. A. Silver parade-helmet any the moss of
Thorsbjerg, Denmark §2
B. Silver bowl imitating glasswork, from
Haagerup, Fyen : : 52
. Shirt of mail from the Vimose, Fyen . 53
. Early Roman silver coins from Denmark. 68
. Silver dish from the Hildesheim hoard,
Hanover 69
. Silver dish from oseridecieen eee 84
. Silver tray from the Hildesheim hoard 84-5
 PLATES bel

XU. A. Woollen trousers from the moss of

Thorsbjerg, Denmark 84-5
B. Glass beaker, probably of Cologne:manu-
facture, fren Nordrup, Zealand : 84-5
AIV. A. Painted glass cup of ‘Nordrup-Jesendorf’
type, from Himlinggje, Denmark. 85
B. Cup of green glass with ribbed decora-
tion, from Vedsted, Zealand . 85
XV. A. Green glass beakers from @rslev Tale
skov, Zealand, and Killerup, Fyen. I0O
B. Green glass drinking-horn from Laerken-
feldt, Denmark ; IOO

. A. Ancient irrigation-canal, Fezzan IOI

B. Pillared glass bowl from the Fezzan IOL

. Mausoleum at Germa, Fezzan 108

A. Amphora from a burial in the Fezzan 109
B. Roman lamp from the ‘palace’ of Tin
Hinan, in the Hoggar 109
. Indian ivory statuette from Pompeii . 132
. Roman and Indian coins from a hoard found
at Eyyal, near Trichur, in Cochin State,
South India 133
XXI. Site of Indo-Roman aime at
Arikamedu, near Pondicherry, South
India . : : , ; 140
XXII. A. Arikamedu: brick foundations project-
ing from river-bank : d : 140-1
B. Arikamedu: excavations in progress,
1945 : : ; I40-I
XXIII. Arikamedu: fragments of Roman amphorae 140-I
XXIV. Arikamedu: fragments of Arretine dishes
and Roman lamp : : I4I
XXV. Bronze statuette of Poseidon, from Kolbapan
western India 148
XXVI. Aand B. Handle of bronze jug pes one
Baroda, western India . rf , 149
Xi ROME BEYOND THE IMPERIAL FRONTIERS

c. Cameo from Karvan, Baroda, western

India 149
XXVIL. a. Bronze jug from Kolhapur, western India 156
B. Silver repoussé emblema representing
Dionysus, from Taxila, Punjab 156
AXVIUL. Terracotta imitations of coins of Tiberius
(obverses), from Kordapur, Hyderabad
State ; . ; : : 156-7
XXIX. Ditto, reverses 156-7
2.5 Air-view of part of eo (irk) Pan
as excavated 157
XXXII. A. Stucco head of an sae oes Taxila 164
B. Stucco head of a satyr, from Taxila . 164
XXXII. A. Stucco head ofa youth, from Taxila . : 164-5
B. Terracotta head based on the Apollo
Belvedere, from Charsada, North-West
Frontier Province . 164-5
XXXII. Stone frieze from Taxila 164-5
XXXIV. Stone relief showing Laocoén and the Trojan
Horse, from the North-West Frontier
Province. Photo, British Museum 165
XXXV. A. Bronze steelyard weight, from the ee
hoard, Afghanistan 168
B. Bronze statuette of eeu wearing he
calathus, from the Begram hoard 168
XXXVI. Bronze figurine of armed horseman, from
the Begram hoard 168-9
XXXVIL. Roman glass beaker with painted decora-
tion, from the Begram hoard 168-9
AXXVIII. Stucco emblemata representing the grape-
harvest and Athena, from the Begram
hoard 169
Map The Roman Empire in the 2nd century A.D.
and Asian trade-routes at end
 I> INTRODUCTION

HIS book is concerned with Roman adventuring

beyond the outermost boundaries of the Roman
Empire. The subject is not new; its literature indeed,
both ancient and modern, is immense; but there are new
examples of it, and old material has recently been reviewed.
In this adventuring Greek and Arab enterprise are wedded to
the Roman faculty for exploitation. The boundaries of the
Empire, particularly in the East, were sufficiently fluid to
ensure a constant awareness of more distant horizons, of
greater riches, more marvels, fresh menaces. A century and
a quarter before Trajan’s thrust to the further bank of the
Tigris, Maecenas at Rome, as Horace tells us, was ‘fearing
what the Seres [Chinese or Central Asians] may be plotting,
and Bactra once ruled by Cyrus, and the discordant tribes
on the banks of Tanais [Jaxartes]’. Nor was this a merely
academic fear. “The credit of the Roman money-market’,
said Cicero, ‘is intimately bound up with the prosperity of
Asia. A disaster cannot occur there without shaking our
credit to its foundations.’ The term ‘Asia’ had to Cicero a
limited implication, but the dictum was broadly true. On
the European frontiers the commercial element was, from
the Roman standpoint, less insistent; nevertheless, un-
conquered Germany remained a recurrent source of anxiety
or, at its mildest, a stimulus to Roman interest and curiosity.
At the outset the word ‘Roman’ requires definition. It
is not indeed closely definable, particularly when applied to
trade. The Italian was no great sailor, just as he was no
great horseman, and not a little of the trafficking beyond the
frontiers involved the use of ships. Therein both fo’c’sle and
quarterdeck must, more often than not, have been manned
by Frisians, Greeks, Levantines, Arabs and others who, like
Kipling’s Parnesius, had ‘never seen Rome except in a
picture’. For the present purpose all these folk will pass
2 ROME BEYOND THE IMPERIAL FRONTIERS

muster as ‘Romans’, so be it that they were directly in the

service of Imperial commerce. There shall be no great
pedantry here in the matter of race or colour or even citizen-
ship.
And as to geography, the widest Imperial limits of the 2nd
century A.D. are assumed as our base-line (end map). Else, for
example, we had need to remind the reader of the Roman
traders who in the 1st centuries B.c. and A.D. preceded the flag
in Gallia Comata or Belgic Britain, with tiresome argument
as to Cunobelin’s foreign moneyers and the extent of his own
part in the cross-Channel trade that littered his squalid huts
at Colchester with fine Arretine dishes or brought silver
eoblets to pre-Claudian Welwyn.
Nor are we concerned here with the sporadic pervasion of
goods from Roman Britain into the outlands of Scotland or
Ireland. Scraps of pottery and occasionally more spectacular
wares reached the native homesteads of Scotland,: and are
there useful mainly for dating undemonstrative local
cultures. A more restricted scatter in Ireland occasionally
has a similar use; as at Tara, where Professor Se4n O Riordain
has dug up sherds of 3rd-century Roman vessels, brought by
traders or (more likely) raiders to this famous seat of kings.?
The bulk of our matter falls into two zones, the European
and the Asiatic. A third is beginning to emerge in the
deserts of North Africa, where there is accumulating evidence
that the traders who brought ivory northwards across the
Sahara to the coastal cities of Libya carried back Roman
goods far into the interior; and though this fact may not be
one of far-reaching historical significance, the importations
have at least a value archaeologically as a potential means of
dating associated African cultures. The East African coast,
1 For an account of these strays, as known in 1932, see J. Curle, ‘An In-
ventory of Objects of Roman and Provincial Roman Origin found on Sites in
Scotland not definitely associated with Roman Constructions’, Proceedings
of the Society ofAntiquaries ofScotland, LXVI (1932), 277-397.
® For Roman goods found in Ireland prior to 1948, see $. P. O Riorddin
in Proceedings of the Royal Irish Academy (Section C), LI (1948), 35-82. The
sherds from Tara were found in 1952-3.
 INTRODUCTION 3

another likely region for Roman contacts, is at present

virtually unexplored.
Our evidence in the following pages will be of two main
kinds. First there are extensive and largely familiar refer-
ences to this trans-frontier venturing in the classical and
Asian records. Secondly, there are numerous discoveries of
Roman or Graeco-Roman commodities, or of native goods
reflecting classical influences, in lands without the Empire.
Most, though not all, of the literary references have a
commercial context. The material ‘finds’, whether in Free
Germany or in further Asia, are often more difficult to
interpret. A majority of them, especially in India or Indo-
China, are manifestly the products of trade or at least of
commercial drift. A minority may be explained as the
fruits of war; as when the three legions under Varus were
overwhelmed by the German tribesmen in a.p. 9 and lost
their whole vast equipment to the victors. Other objects of
value found their way across the borders as propitiatory
gifts to native princes; as in our own time golden sovereigns,
for example, have flowed diplomatically into Arabia.
Again, tribal movement across or along the frontiers of
Central Europe must have involved a displacement of Roman
things on a fairly considerable scale. By and large, the
problem of interpretation is not a simple one. It must be
discussed at later stages in the light of individual circum-
stance.
But, when all is said, trade will be discovered as the
dominant factor in these remote contacts. The Roman
traders and their agents were venturesome folk, whether we
regard the enterprising knight who in the time of Nero made
his way to the Baltic and brought back a great load of
amber, or the pioneers who in the 2nd century reopened the
trans-Asian ‘Silk Route’ beyond Tashkurgan, or again those
others who in A.D. 166 carried gifts to the emperor of China
and so opened or reopened a sea-way to the Far East.
Amber, pearls, pepper, silk, such were the goals of all this
trafficking, and most of it is of interest to us to-day in the
4 ROME BEYOND THE IMPERIAL FRONTIERS

sense that perilous voyaging and travellers’ tales are of

perennial interest to those who care for human enterprise.
But once at least this trafficking did something more than
merely scatter Roman bric-d-brac across the world. In
Afghanistan and what is now West Pakistan it was the
instrument of a cultural contact which had a far-reaching
effect upon the history of art.
It stands to reason that the surviving vestiges of this
interchange are in kind but a fraction of the whole, and
that their survival is fraught with all manner of accident.
The slaves and hunting-dogs and corn imported into the
Roman world from Britain; the cotton cloth from India;
the wine that flowed into Free Germany or was borne by the
monsoon to the tables of Indian princes; the ‘Seric’ skins from
High Asia, the silk from China; the spice that filled the
Pepper Barns of Rome; for these and much else we have
to rely almost entirely upon chance scraps of history. In
the circumstances, the surprising thing is not how little but
how much has survived in the form of material evidence.
Roman coinage reached India in quantity and even pene-
trated to Indo-China. Roman glass has been recorded
from the Trondheim fiord on the one hand and (admittedly
with less certainty) from Honan and Korea on the other.
Silver cups from Italy and the provinces equipped the graves
of the native nobility in Germany, Czechoslovakia, Poland
and Denmark. Bronze vessels reached Norway and western
India. To western India again, and to Afghanistan, came
Roman bronze figurines, and others have been recovered
from the native homestead-mounds of the Low Countries.
Red-glazed pottery from Italy or the western provinces is
found alike in Poland and on the Coromandel Coast.
Roman arms and armour were dedicated in the meres of
Jutland. The list might be extended, and will in due course
be dealt with categorically. Meanwhile, enough has been
said to indicate its range.
From this point, the book will be divided into three
parts, dealing respectively with Europe, Africa and Asia.
 INTRODUCTION 5

Europe will for our purpose comprise Free Germany (in-

cluding Scandinavia); Africa will be subdivided into the
Sahara and the East Coast; Asia will include the Indo-
Pakistan sub-continent and Afghanistan on the one hand, and
the Far East on the other.
 PART I - EUROPE
x ke *

Ne TREE GERMANY
TrESVIVERARY EVIDENCE
EFORE the middle of the rst century B.c. hardy
B Roman business-men were already operating amidst
the German tribes across the Rhine. ‘The Suebi’,
Caesar tells us (De Bello Gallico, IV, 1-3), ‘by far the largest and
most warlike nation amongst the Germans, give access to
traders, rather to secure buyers for what they have captured
in war than to satisfy any craving for imports.’ The Suebi
were an amalgam of German tribes who, in the time of
Strabo (late 1st century B.c.) were living between the Rhine
and the Elbe, and in that of Tacitus, about a.p. 98, were
occupying ‘more than half Germany’; more precisely,
north-western Germany and the middle and lower Rhine as
far as the sea. Caesar adds that ‘they suffer no importation
of wine whatsoever, believing that men are thereby rendered
too soft and womanish for the endurance of hardship’. This
remark implies that elsewhere wine was an acceptable im-
port; Tacitus (Germania, 23) states that ‘the Germans who
live nearest the Rhine can actually get wine in the market’,
whilst those in the interior “extract a juice from barley or
erain, which is fermented to make something not unlike
wine’. Roman wine-vessels from the Jutland Peninsula are
found to have contained a fermented drink made from malt
and berry-juice, recalling the cranberry wine, mixed with
myrtle and honey, of which desiccated remains are found
already in Danish burials of the Bronze Age, over 1,000 years
eatlier. Caesar does not tell us what the Roman traders
gave in exchange for surplus Suebian loot, but it may be
supposed that silver money and Roman metalwork formed,
now as later, a part of the apparatus.
8 ROME BEYOND THE IMPERIAL FRONTIERS

Certainly by the rst century A.D. Roman silver coinage

was an established factor in the German trade. Again
Tacitus is our historian (Germ. 5): “The Germans nearest us
value gold and silver for their use in trade, and recognize
and prefer certain types of Roman money. (The peoples of
the interior, truer to the plain old ways, employ barter.)
They like money that is old and familiar, denarii with the
notched edge and the type of two-horse chariot. Another
point is that they try to get silver in preference to gold.
They have no predilection for the metal, but find plenty of
silver change more serviceable in buying cheap and common
goods.’ The distribution-map (fig. 6) shows that by the
3rd century Roman coinage had spread into Germany far
beyond the vicinity of the limes. The preference for the
older types of silver obviously derives from the dilution of
the denarius by Nero, and reflects the native use of the
money, not as a token currency backed by a state guarantee
(for there could be none such beyond the frontier), but as so
much bullion of honest intrinsic worth. That intrinsic
worth was permanently impaired by the monetary ‘reform’
of a.D. 63. More will be said of this matter in a later
chapter (p. 63).
The early trade across the frontiers was not without its
dangers to the adventurers who undertook it. Some thirty
years after Caesar's casual reference to Roman. traders
amongst the Suebi, the legate of Augustus ‘took vengeance
on certain Celts [i.e. Germans] because they had arrested and
slain Romans who entered their country to trade with
them’ (Dio, LIII, 26). Such vicissitudes were doubtless a
part of the day’s work. On the whole, however, it is to be
assumed that this trafficking was welcome enough on both
sides. Thus in a.D. 18, as Tacitus tells us (Annals, II, 62),
there were domiciled at the capital of the king of the Mar-
comanni, i.e. in Bohemia, a number of sutlers and traders
‘implanted first on foreign soil by commercial privileges,
then by the lure of increased profits, and finally by oblivion
of their country’. The king, Maroboduus, had himself
 FREE GERMANY: THE LITERARY EVIDENCE 9

been to Italy in his earlier days and doubtless encouraged this

trade as a matter of policy and as a demonstration of en-
lightenment. Tacitus is reinforced by the distribution-maps
(e.g. fig. 7), which show in the earlier part of the 1st century,
but rarely afterwards, a concentration of Roman relics in the
Marcomannian area. This area was approached most readily
from the neighbouring ‘amber route’, which left the Danube
at Carnuntum and penetrated up the tributary valley of the
March towards the German plain and the Baltic.
The route to the Baltic is itself the subject of a well-known
episode recounted by Pliny (Natural History, XXXVII, 45).
In the time of Nero a Roman knight, agent of a certain
Julianus, made the arduous journey to the Baltic coast,
where he visited commercia or agencies, and eventually
returned, presumably by way of East Prussia and Poland,
with a great quantity of amber. The pioncering character
of this journey implies that previously the northern sectors of
the amber traffic had been in unrestricted native hands, and
the knight, though primarily concerned, it seems, with the
equipment of gladiatorial shows for Nero, may well have
been prospecting incidentally with a view to simplifying
the trade and reducing its costs. We would give much for a
sight of his report.
Commercial enterprise was not all one way. Tacitus
(Germ. 41) states that the tribe of the Hermunduri, who lay to
the north of the upper Danube as far as Thuringia, ‘are the
only Germans who trade with us, not only on the river-
bank, but deep inside our lines, in the brilliant colony that is
the capital of Raetia. They come over where they will,
and without a guard. To other nations we only show off
our arms and our camps; to them we expose our palaces and
our country-mansions—and they do not covet them.’
Free trade of this kind may not in fact have been quite as rare
as Tacitus affirms. Much later, after the defeat of the
insurgent Marcomanni in A.D. 173, Marcus Aurelius ‘estab-
lished the places and the days for their trading, for these had
XII, 15). Whatever the
not been previously fixed’ (Dio, LX
10 ROME BEYOND THE IMPERIAL FRONTIERS

custom on the Rhine, traffic across the Danube frontier

would appear to have been substantially unregulated until
the disturbances of the latter half of the 2nd century.
Trade apart, Roman things reached Germany through
war and through diplomacy. For the latter two references
from the Germania (5 and 42) are in point: “One may see
among them [the Germans] silver vessels which have been
given as presents to their envoys and chiefs’; and again, ‘the
might and power of the kings [of the Marcomanni and the
Quadi] depend upon the authority of Rome. These kings
occasionally receive our armed assistance, more often our
financial, and it is equally effective’. It may be assumed
indeed that diplomatic bribery was a normal feature of the
relationships between Rome and Free Germany, and helped
very materially in the scattering of coinage and other
valuables. Thus when, about A.D. 90, Chariomerus, pro-
Roman king of the Cherusci, was driven from his kingdom
by the Chatti, he appealed to Domitian and “did not secure
any military aid but received money’ (Dio, LXVII, 5). And
Decebalus, king of the Dacians, was bought off both by
lavish payments and also by the receipt of technical aid.
About a.p. 89 Domitian sought to keep him quiet by
sending ‘large sums of money and artisans of every trade,
both peaceful and warlike, and promised to keep on giving
large sums in future’ (Dio, LXVI, 7). This subvention was
continued for 10 years or more, and one of the reasons for
Trajan’s campaign against the Dacians was his anxiety ‘at
the amount of money they were receiving annually’ (Dio,
LXVIII, 15). The submission of the Dacian king in a.p.
102 included the surrender of his ‘arms, engines and engine-
makers’ (Dio, LX VIII, 9), presumably the Roman technicians
seconded to him by Domitian. But the flow of coinage
across the frontier increased rather than diminished as time
went on. The Roman gold which was freely distributed in
barbarian Europe in the time of the Goths and, above all, the
Huns reflects in part the continuance of this policy on an
exaggerated scale at the lower limit of our period.
 IID PREE GERMANY
ROUTES AND MARKETS

Be we pass from the historical to the material

evidence for this interchange, something may be said
of the principal areas and routes with which our study
is concerned. To do this at so early a stage is of course to
reverse the logical sequence, and to anticipate the map-
patterns which are found to emerge from the inquiry. But
the active work of Scandinavian and German scholars
during the past few decades has rendered this inverted
procedure possible, and it is justifiable on the ground that a
summary survey such as the present is more readily in-
telligible if the geographical aspects are sketched in at the
outset.
First let it be said that no special attention is given here
to the occurrence of Roman things within a belt fifty miles
wide outside the formal frontier or limes. A marginal spread
of this kind speaks for itself and can here be simultaneously
defined and dismissed. In the time of Hadrian, the frontier
on which it was based extended up the Rhine nearly to
Coblenz, then bore eastwards in a great salient round the
Taunus, southwards into Wurttemberg, and eastwards
again to the Danube above Regensburg. Thence it followed
the Danube to the Black Sea, save where it bent northwards
to the Carpathians to enclose the province of Dacia. This
broad frontier-zone across Europe frames on the south and
west the Free Germany with which the present chapter deals.
In the north and east, that Germany is regarded as extending
to Norway, Sweden, Latvia, Poland and South Russia,
without any attempted discrimination of racial and linguistic
boundaries, which must indeed have been numerous and
sometimes formidable. The whole vast region was an
officina gentium, a welter of nations increasingly mobile as
12 ROME BEYOND THE IMPERIAL FRONTIERS

the Imperial era wore on, already sufficiently kaleidoscopic

before the mongoloid Huns brought final chaos into it at the
end of the 4th century. In the circumstances it need scarcely
be emphasized that the term ‘free’ is here merely a euphem-
ism for ‘outside the formal boundaries of the Roman Empire
in the 2nd century A.D.’, ic. the boundaries defined above,
always with the reservation that those boundaries were
extensively withdrawn by the rupture of the Rhine-Danube
limes in A.D. 258-260 and the surrender of Dacia beyond the
Danube a dozen years later.

ROUTES

Far beyond the range of mere frontier-diffusion a number

of main routes carried goods to and from the depths of this
Free Germany (fig. 1), and the arterial traffic was supplemented
by lateral pervasions which tend to blur the map-pattern and
complicate its interpretation. Within the Empire the traffic
was focused on three great regions: Italy, Gaul with the
German provinces, and the periphery of the Black Sea. Of
these, the first to dominate the scene was naturally Italy;
but in the 2nd century a.p. the developing industries of Gaul
and the Rhineland were of increasing importance, and in the
3rd century probably outstripped Italy in their trans-frontier
influence. Shortly after the middle of that century, the
tribal confederacy known as the Goths, who had long been
moving southwards from their Baltic homeland, encom-
passed the northern and western shores of the Black Sea and
now became an instrument in the diffusion of the rich
craftsmanship of Byzantium, South Russia and the Near
East towards barbarian Europe. Trade, conquest, sub-
vention, drift, all contributed to the complex process of
dispersal and are usually hard to discriminate. But in one
way and another, by the 3rd and 4th centuries Gaul and the
lands of the Black Sea had between them largely replaced
Italy as the sources of classical craftsmanship in Free Germany,
whilst subsequently the Eastern Empire of Byzantium
 FREE GERMANY: ROUTES AND MARKETS 13

became the major factor. Ofall thiscomplex trade and move-

ment through four centuries, the principal lines of intercourse

2 Heights over
* 500 Metres

100 200 300 400 500 K;Jometres

Fig. 1 Principal European trade-routes

and the principal termini are tolerably clear; only the

motives and occasions are, more often than not, held from us.
14 ROME BEYOND THE IMPERIAL FRONTIERS

From Italy in the first two centuries a.p. came, above all,
bronze wares and glass. Of both, Campania was probably
the foremost producer, although the accidental preponder-
ance of material from the Campanian cities of Herculaneum
and Pompeii tends to distort our perspective. At any rate,
Aquileia at the head of the Adriatic was the chief point of
assembly; it was inevitably Aquileia, for instance, that
became the goal of the insurgent Marcomanni and their allies
when they burst through the Julian Alps in a.p. 167. No
doubt the tribesmen followed well-trodden trade-routes
between the Danube and Aquileia, in particular the high-
road built by Augustus through the province of Noricum
(Austria) and Upper Pannonia to the Danubian fortress of
Carnuntum, placed strategically at the junction with the
March below Vienna. Noricum itself developed metal
industries with local characters which can be recognized in
eraves far away in Jutland, and was clearly not blind to the
strategic possibilities of its position.
From Carnuntum an early route led northwards up the
valley of the March, forking westwards into Bohemia or
continuing northwards through the Moravian Gates into
the German plain. There it struck the upper reaches of the
Oder, but soon branched northwards again towards Kalisz
in western Poland and so reached the lower Vistula and the
Amber Coast of the Baltic, which was its main objective.
Traffic along this ancient route—indeed, long-distance
traffic of the kind in any part of Free Germany—must have
been a matter of constant relay with accumulating dues, and
it has been suggested above that the exploratory journey of
the knight of Julianus in Nero’s reign was in part an attempt
to simplify and cheapen the process. By whatever means,
throughout the period with which we are dealing Roman
things circulated fairly constantly through this natural artery
from the lands tapped by the Danube if not always from
Italy itself Roman coins of the and century reached the
lips of the dead in the neighbourhood of Konigsberg, bronze
vessels of the 2nd and 3rd centuries were entombed in the
 FREE GERMANY: ROUTES AND MARKETS 15
lands south and west of Danzig Bay, and in the 4th and sth
centuries Roman gold bestrewed the Amber Coast and the
Swedish islands as token of the new, uneasy partnership of
barbarism with the civilized world.
* The western branch-route from the March into Bohemia
had only an intermittent importance, but was more in-
strumental than any other known route in opening up
central Europe to Roman trade in the early days of the
Empire. The historical position of Bohemia, home of the
powerful Marcomanni and their neighbours the Quadi, is
sufficiently outstanding to demand a short historical excursus
at a later stage. Meanwhile, it will suffice to observe that in
the early decades of the 1st century A.D. Bohemia was a
leading entrepdt of Roman trade in Free Germany and must
have been instrumental in the diffusion of Roman goods
northwards into Prussia and beyond; that by the middle of
the century this trade had diminished to vanishing point;
but that some sort of revival took place under the later
Empire, when coinage and bronze-wares again reached the
region in modest quantities. Bohemia and Silesia contain
natural resources, notably tin, of value to the classical world,
but the independent spirit of the local tribesmen in their
secluded and forested highland homes was an obstruction to
regular interchange.
Another approach into and through Bohemia from the
Danube valley lay from the vicinity of Linz across to the
Moldau and via Prague, Dresden, Meissen, Leipzig and Halle
to the Elbe, the German Plain, the Jutland peninsula and
Scandinavia. The route was, in part at any rate, an old one,
but as a whole its importance in the Roman period appears
to represent a secondary development and never to have
rivalled that of more easterly and more westerly alternatives.
In the west, sea-routes and land-routes combined to
supply western Germany and the Baltic lands. The sea-
routes are necessarily less sensitive to archaeological recogni-
tion but must not on that account be underrated. In the
proud record of Augustus at Ankara it is proclaimed that
16 ROME BEYOND THE IMPERIAL FRONTIERS

‘My fleet made an ocean voyage from the mouth of the

Rhine to the eastward as far as the boundaries of the Cimbri
whither before that time no Roman had penetrated either by
land or by sea; and the Cimbri, the Charydes, the Semnones
and other German peoples of that same region sent emissaries
to seek my friendship and that of the Roman People’.
These were tribes of the Elbe valley and the Jutland peninsula,
though whether the fleet circumnavigated the peninsula on
this occasion may be doubted. The Romans were timid
sailors, and it may be supposed that the longshore tribes,
above all the Frisians of the Dutch coast, became the principal
carriers of the subsequent sea-borne trade hereabouts. Their
main markets towards the north-east lay in Jutland, where
amber was procurable, and in the populous Danish islands;
and a short land-route across the base of the peninsula
between the North Sea and the western Baltic may be
presumed. Trans-peninsular routes were a normal sub-
stitute for circumnavigation in ancient times, and a more
ambitious example within the Ist century A.D. will emerge
when we review the evidence from South India (p. 144).
By land, the eastern tributaries of the Rhine offered
access from the Roman frontier to Westphalia and Thuringia.
The Lippe, with the fortress of Vetera (Xanten) at its base,
the Ruhr from Asciburgium, less certainly the Sieg from
Bonn, all played some part in this traffic; above all, perhaps
the Lahn and the Main, the latter overlooked by the key-
fortress at Mainz. These valley-routes tended to converge
upon Paderborn in eastern Westphalia and the Weser-
crossings at Minden, Hameln and Héxter, whence dispersal
into the broad lowlands towards the Elbe and even to the
Baltic was not difficult.
South of Mainz the country east of the frontier becomes
more broken and rugged, and clear traces of arterial traffic
cease. Nevertheless, the province of Raetia, overlapping
the upper waters of the Danube, must have acted in some
measure as a link between Upper Germany and Gaul on the
one hand and the Danubian axis on the other, and there are
 FREE GERMANY: ROUTES AND MARKETS 17

hints that pottery (terra sigillata and ‘Rhenish’ ware) and other
objects from western factories occasionally passed this way.
But the main commercial activity of Raetia was directed
historically towards the north. There in the 1st century
A.D. lay the considerable nation of the Hermunduri, whose
whereabouts at various periods has given rise to considerable
argument! and doubtless, indeed, fluctuated with the
characteristic mobility of the German tribesmen. They are
strangely ignored by the geographer Ptolemy about a.p.
150, but were placed by Strabo (end of the 1st century B.c.)
east of the Elbe. On the other hand, between 7 and 3 B.c.
Ahenobarbus, commander on the upper Danube, had
penetrated as far as the Elbe or the Saale and, in doing so,
had (according to Dio, LV, toa) ‘intercepted the Hermun-
duri, a tribe which for some reason or other had left their
own land and were wandering about in search of another’.
He settled them on territory vacated by the Marcomanni
when they moved from the neighbourhood of the Main and
the Neckar to Bohemia (see below, p. 19). The name of the
Hermunduri has been connected philologically with Thur-
ingia, but, though they appear to have occupied that region
during some part of their history, the identification cannot
be sustained. Tacitus, as we have seen, places them nearly
100 miles further south, describing them as ‘the only Germans
who trade with us . . . deep inside our lines, in the brilliant
colony that is the capital of Raetia’ (p. 9). We may
perhaps recognize a northern and a southern branch of the
tribe, one in Thuringia and the Saale-Elbe valley, the other,
presumably coterminous, in Bavaria round about Nurem-
berg. Archaeology emphasizes the commercial importance
of Thuringia, beside a traditional route between the Baltic
and the west, but offers no clear illustration of the statement
of Tacitus in regard to Raetia. Its reticence in this matter
is a sufficient reminder of the incompleteness of our picture
in times or regions where history fails us.
If we turn eastwards again, to modern Hungary or ancient
1 Reviewed in Germania, XXIII (1939), 262ff.
18 ROME BEYOND THE IMPERIAL FRONTIERS

Dacia and as far east as South Russia and the environs of the
Black Sea, we find ourselves in lands which have been in-
adequately explored and are at the present time inaccessible.
East of Budapest the valley of the Theiss marked a route
northwards from the Danube to the Carpathian passes and
so into Galicia and central Poland and on to the Baltic—a
route which may be supposed to have shared in the amber
trade but can scarcely have depended wholly upon it. Then,
still further east, the rich environs of the Black Sea, important
to us in the later 3rd and 4th centuries, were tapped by the
valleys of the Dniester and the Dnieper, the former pointing
to Galicia and the Vistula or the Oder, the latter to White
Russia and the south-eastern Baltic. There can be little
doubt that with ampler knowledge these eastern routes would
be found to rival the better-known arteries of the west; the
heavy distribution-pattern of 4th- and sth-century finds,
mostly from the Eastern Empire, on the Swedish islands and
the East Prussian coast is sufficiently significant.

MARKETS

From routes we turn to the markets which they supplied,

and in doing so encounter at once the initial handicap to
which reference has been made above: the vicissitudes of
survival. Nevertheless, it seems likely that the principal
regions of concentration on our maps are reasonably re-
presentative, if by no means complete. Of the somewhat
left-handed trade which, as Caesar tells us (p. 7), the Suebi
of north-western Germany encouraged in the 1st century
B.C., there is no recognizable vestige unless certain Italic or
Italo-Gaulish bronze vessels of that general period found in
cremation-burials of the Weser-Elbe zone (Eggers, map 3; see
Bibliography, p. 182) are attributed to it. For a majority of
these vessels, however, an approach from the south-east via the
upper valley of the Oder is more likely. On the other hand,
Bohemia, land of the Marcomanni and the Quadi, comes ap-
propriately on to the map during the early decades of the ist
 FREE GERMANY: ROUTES AND MARKETS 19

century A.D., in historical circumstances which may be recalled.

Bohemia. ‘Next to the Hermunduri dwell the Naristi,
followed by the Marcomanni and the Quadi. The Mar-
comanni are conspicuous in renown and power; they won
the very land they now hold by their bravery, when they
drove out the Boii.... The might and power of the kings
depend upon the authority of Rome. These kings occasion-
ally received our armed assistance, more often our financial,
and it is equally effective.’ So Tacitus (Germ. 42). It would
appear that the Marcomanni were originally settled in the
valleys of the Main and Neckar, whence in the time of
Caesar they took arms with his opponent Ariovistus; but in
9 B.c. Drusus, a few weeks before his death, launched an
attack upon them, with the important consequence that
they migrated eastwards to the uplands of Bohemia. There,
in country at the same time fertile and sufficiently difficult
to discourage interference from the west or south, they
occupied the old home of the Boii and quickly developed
into a dominant native power.
Their leader in this phase of migration and aggrandisement
was the celebrated king Maroboduus, who knew Rome and
may actually have served with the Roman army. His
knowledge enabled him both to deal with Roman diplomacy
and to apply his Roman training to the organization and
defence of his own large kingdom. His strength and
ambition were not underrated by his opponents. A careful
and skilful policy of encirclement culminated in a.p. 4 in a
threefold thrust into his country—from Mainz, from Raetia
and from Illyricum—and his last-moment escape was due
only to a timely revolt in Ilyricum behind the advance.
The frustrate campaign ended in a treaty whereby Marobo-
duus was acclaimed a friend of the Roman people. But,
where Rome had failed to tread, Maroboduus’s fellow
Germans stepped in and won the day. From the north,
Bohemia is penetrated by the upper valley of the Elbe and
its tributary the Moldau. Up these valleys in a.p. 18
streamed a host of rival tribes under the leadership of
20 ROME BEYOND THE IMPERIAL FRONTIERS

Arminius who, since his grim victory over Varus in A.D. 9,

had been the champion of German freedom. The threat
was more than even the power and prestige of Maroboduus
could withstand; and when Arminius received support from
an exiled Marcomannian, a certain Catualda, bringing at
least diplomatic encouragement from Rome, the king himself
fled to his double-dealing Roman friends, to end his days,
years later, in sanctuary at Ravenna.
Tacitus, it will be recalled (p. 8), tells us that, when
Catualda entered the capital of the vanquished Maroboduus,
he found a number of Roman traders in residence there,
operating under commercial privileges which had pre-
sumably been included in the treaty of a.p. 4. That treaty
may be supposed to have regularized an existing state of
affairs rather than to have initiated a new one. Roman
traders, like those of ascendant powers in later ages, were
ever wont to precede the flag, and the Roman training of
Maroboduus must have inclined him to encourage them,
with or without a formal writ. Be that as it may, Italic
bronze vessels of the first half of the 1st century A.D. are at
home in Bohemia (fig. 7), together with a considerable
coinage, mainly of silver, but including gold; and from
Bohemia these imports spread northwards by the river
valleys to the North Sea and the Baltic.
There is evidence that at this early period the Marcomanni
played something of a creative role in the development of
central European craftsmanship. Geographically and ethnic-
ally they were indeed in a favourable position to do so.
Coming from the west, they were in contact with the Celtic
world; they looked northwards to the German plain and
Scandinavia; partially romanized Noricum (Austria) was at
their doorstep; Roman agents were in their midst. They
fell thus under a number of diverse stimuli and for a brief
time formed a sort of cultural exchange and mart.t. Thus
1 See a discussion by O. Almgren, ‘Zur Bedeutung des Markomannen-
reiches in BOhmen fiir die Entwicklung der germanischen Industrie in das friihen
Kaizerzeit’, in Mannus, V (Wiirzburg, 1913), 255-78.
Silver
, cup,P, fpartiallyyggilt, from a ‘chieftain’s’ gcrave at Dollerup,F Denmark.
z (See p. 40)
 I]

Bronze JUS
jug from a ‘chieftain’s’ Sgrave at Hoby,) Laaland. ex|to
(See p. 37)
 FREE GERMANY: ROUTES AND MARKETS 21
certain types of brooch and belt-fittings, particularly with
stylized animal-heads, are common to Bohemia and
Noricum. On the other hand, a well-known type of
brooch with two piercings or circles (‘eyes’) on the head of
the bow—a late La Téne type which in subsequent forms
was to become characteristic of the Rhineland—came in
from the opposite direction, from Brandenburg, Posen and
Saxony, and was developed in Bohemia, whence it passed
to the Rhine after the collapse of the Maroboduus régime.
And it was seemingly in Bohemia that many of the earlier
drinking-horns of this period were made, to judge from the
fact that there, within a restricted territory, occur all the
earliest known types of bronze fittings which are elsewhere
scattered sporadically from the Rhine to Gotland. It was an
established custom amongst the Germans to supply their
drinking-horns with metal rims and terminals and to equip
them with chains or straps for suspension; in the words of
Caesar (B.G. VI, 28): “They zealously collect the horns and
encase the edges with silver, and then at their grandest
banquets use them as drinking-cups.’ Examples will be
mentioned later (pp. 36, 39, 42).
After the exile of Maroboduus, Italo-Bohemian commerce
rapidly deteriorated. The latest Roman gold coin from
Bohemia during the 1st century is a solitary aureus of Titus
(A.D. 79-81), and the general run of imports had ceased
nearly half a century before that date. The consolidation of
Roman authority in southern and western Europe had
opened up alternative routes and markets, and the Bohemian
tribes, increasingly isolated and (it may be) overcrowded,
were becoming restive. An umeasy peace, interrupted
momentarily at the time of Domitian’s Dacian war by a firm
Roman diplomacy, itself bordering upon war, ended in
A.D. 167, when the Marcomannian king Ballomar led his
tribesmen and allies across the Danube, defeated a Roman
army, and besieged the pivotal market-town of Aquileia on
the Adriatic. The Roman emperor reacted with determina-
tion; the invaders were beaten back, and in 172 Marcus
Cc
22 ROME BEYOND THE IMPERIAL FRONTIERS

Aurelius led his forces to victory against the Quadi, on the

eastern flank of the Marcomanni. In the following year
came the turn of the Marcomanni themselves. Their
territory was occupied and a part of the population was
transported.!. Thereafter they never again so nearly threat-
ened the safety of Rome, but they did not cease to contribute
to the recurrent restlessness of the Danube frontier, and such
trade as lingered on was canalized and constrained by Roman
treaty.2 Only in the latter part of the 4th century did gold
coinage and bronze vessels again find their way to Bohemia
in appreciable quantity, at a time when eastern Europe was
in the grip of the Huns.
Thuringia. Between the Thuringian Forest and the Elbe,
in a rough triangle formed by Erfurt, Leipzig and Magdeburg
and extending into the Forest, a concentration of coins,
particularly of the Middle Empire, is less easy to explain.
With them are vessels of bronze, pottery and glass, the two
last certainly from the west. They presumably arrived in
part by the river-routes from the Rhineland (p. 16), above
all perhaps by an old way from Mainz over the Wetterau,
continued in and beyond Thuringia by the valleys of the
Saale and the Elbe. This was, incidentally, the line of a
prehistoric amber-route from Jutland to the middle Rhine;
a route which, until chaos supervened in the latter half of
the 3rd century a.D., alternatively continued south-eastwards
up the Elbe to Bohemia and the Danube, and so connected
Thuringia with Noricum and the eastern traffic-system.
Furthermore, if the Hermunduri extended to Thuringia,
we have the clear witness of Tacitus (p. 9) for active inter-
course more directly with the south, with the province of
Raetia. Thuringia would thus appear to have occupied
something of a nodal point in European trade. It is at least
certain that, in a fertile region accessible though at the same
1 Dio, LXXII, II, 4, refers to a revolt of Marcomanni settled at Ravenna,
where, incidentally, Maroboduus had been given sanctuary over a century and
a half previously (see above). After the revolt, Marcus evicted the settlers.
2 Marcus ‘established the places and the days for their trading (for these had
not been previously fixed)’. Dio, LXXII, 15.
 FREE GERMANY: ROUTES AND MARKETS 23

time screened by a succession of massifs and woods, the

Thuringians or Hermunduri developed a liking for Roman
things of which some account will be given in later sections.
The Baltic Coast and Eastern Europe. A thickening of
Roman goods in the approaches to the gulf of Danzig (e.g.
fig. 12) draws attention to the geographical importance of
East Prussia and its environs in the commercial and political
picture of the period. Historically our information is of the
scantiest, although the wealth of the Pomeranian-East
Prussian coast in amber, and the popularity of amber in
the ancient world, are established facts. Pliny’s reference,
already cited (p. 9), toa Roman knight who, in the time of
Nero, set out on an exploratory journey from Carnuntum
to the Baltic coast and returned with a great quantity of
amber merely underlines familiar knowledge.
But if history fails us, archaeology has produced some
significant results. The coin-evidence is particularly striking.
A majority of the individual Roman coins from East Prussia
have been found in graves, which have produced upwards of
559 of them (Bolin; see p. 63, footnote 3), mostly sesterces.
By far the greatest number of the coin-graves has been found
north-east of the gulf of Danzig, in the coastland between the
rivers Pregel and Niemen, but they extend southwards to the
old border with Poland and Russia, where their apparent
cessation may be due merely to lack of research. In some
graves the coins were introduced as ornaments, but in others
they had been placed between the lips of the dead, the
Charon’s fee of the classical world. Save in Samland where
more than half of the coins are earlier than A.D. 138, a
large majority of them date from A.D. 138 to 180 (the
accession of the emperor Commodus), and in Samiland itself
of four hoards of sesterces three end with Commodus and
the fourth with Septimius Severus. The second half of the
2nd century was evidently the optimum period.
Before conclusions are drawn from this concentration of
coinage, one overriding factor must be considered: namely,
the circumstances of the preservation of the coins. Be it
24 ROME BEYOND THE IMPERIAL FRONTIERS

repeated that most of them have been found in graves. The

prerequisite condition is that the inhabitants of the region
concerned were in the habit of including coins amongst their
grave-goods. Similar coins may well have circulated just as
freely amongst a neighbouring people whose funerary
fashions were of a different sort. And that such was almost
certainly the case has been demonstrated by Eggers with the
aid of two maps of the Vistula-Baltic region (fig. 2). One
map shows the coin-distribution just recounted; the other
shows the distribution of Roman bronze vessels found in the
graves of the same area. The one map is almost exactly
the complement of the other; where the coin-burials end,
the burials with bronze vessels begin. The inference is not
in doubt. In fact we are confronted, not with the vagaries of
Roman trade, but with those of native burial-custom. The
lesson is one to bear in mind in instances where the materials
for such demonstration may not be available.
To the further obvious question, What ancient frontier—
cultural or political—does this change in custom reflect?
there is no clear answer. The region was basic in the great
movements of Goths, Vandals and Burgundians under the
Later Empire, and, behind these movements, the frontiers of
the 2nd century are too indistinct for use. The cultural
division extending southwards from the gulf of Danzig
roughly along the line of the Passarge river is clear enough
archaeologically, but cannot be translated into history.
In the 3rd and 4th centuries the lands flanking the lower
Vistula became the repository of certain types of bronze-
and glass-ware for which an origin in south-eastern Europe
is predicated. These include bronze bowls covered with
curved flutings, and certain types of glass beaker of kinds not
present in the Rhenish or Belgian factories. More will be
said of these wares at a later stage. Meanwhile, it may be
observed that their distribution in the Baltic lands is rein-
forced by the astonishing concentration hereabouts of gold
coins, solidi, mostly of the Eastern Empire and of dates
ranging from the end of the 4th through the sth century.
 FREE GERMANY: ROUTES AND MARKETS 25

Gold had now replaced silver as the staple currency of

treasure, and gold matched the taste of east-European

GRAVES WITH ROMAN

BRONZE VESSELS

Scale 2 50 199 Yi'les

Fig. 2. Differential distribution of Roman bronze vessels

and coins. (After H. J. Eggers)

barbarism. It lined the Baltic and flowed over into the

Swedish islands and the pages of the sagas.
26 ROME BEYOND THE IMPERIAL FRONTIERS

The new emphasis thus attached to a south-north circula-

tion in eastern Europe is consistent with certain familiar
historic trends under the Later Empire. The Goths, whose
earlier home had perhaps been in that part of southern
Sweden which still bears the name of Gotland or Gétalund,
had by the 2nd century a.D. extended across the Baltic to the
Danzig plain and its environs. At about the time when their
continental neighbours, the Vandals and the Burgundians,
began to thrust south-westwards to the Rhine, they them-
selves turned southwards, advancing in loosely integrated
hordes towards the Danube, the Balkans and the Black Sea.
By the middle of the 3rd century they had become an instant
threat to the Danubian territories of the Empire. Roman
armies, rallied by one emperor or another in this chaotic
time, made a poor show against them, and it was not until
A.D. 269 that the second Claudius earned the title “Gothicus’
by crushing them at the battle of Naissus, far down in
modern Yugoslavia. Meanwhile, a host of them had made
its way to the Crimea, devastated the rich cities round the
shores of the Black Sea, and established a powerful Gothic
kingdom there. In a.D. 276 another emperor, Tacitus, was
earning the same title “Gothicus’ in an attempt to check them
in the heart of Asia Minor.
This folk-wandering over vast stretches of eastern Europe
and western Asia must not, however, be regarded in terms of
an advancing army. Rather was it the intermittent and
partial thrusting of droves, sometimes larger, sometimes
smaller, from an inchoate mass of tribes and septs vaguely
co-ordinated as ‘Goths’, but dependent largely on the
accidents of individual leadership. Long after a Gothic
horde had penetrated to Cilicia, other Goths were still
threatening the Danubian frontier of the Empire or were even
being enrolled in the Roman army; and, as every schoolboy
knows, in A.D. 410 a Gothic king sacked Rome itself. In the
3rd and 4th centuries, from the Baltic to the Danube and the
Black Sea, in contact alike with Italy and the Eastern Empire,
was a continuum of German peoples who may be loosely
 FREE GERMANY: ROUTES AND MARKETS 27

aggregated as Goths: to be complicated but not entirely

broken at the end of our period (after a.p. 370) by the first
incursions of the Asiatic Huns, whose depredations in and
beyond the Danubian regions lie mostly outside the present
story. The context was complete for that increasing
circulation of Roman, and particularly East Roman, things
throughout this wide, uneasy zone and turbulent period.
Our bronzes and glassware of non-western types in East
Prussia and the Baltic are an acceptable illustration of the
process; whilst upwards of 54 hoards of late 2nd or early
3rd century denarii in Russia, containing more than 11,000
coins, and at least 17 hoards of similar kind from Poland
(over 10,000 coins)! probably reflect the same complex of
events. Finally, the Runic alphabet which appears to have
been evolved in contact with the Greek and Latin alphabets
shortly before or after A.D. 200 may have been invented by
the Goths?; it at least owed its early and widespread distribu-
tion between south-eastern Europe and Scandinavia to the
constant interchange of commodities and ideas within this
Gothic zone.
Jutland and the Baltic Islands. Mention has been made in
the preceding section of the abundant gold coinage, mostly
of the Eastern Empire, which in and after the end of the 4th
century A.D. overflowed to the Prussian coast and the
Swedish islands, Gotland and Oland. By the sth century
these islands were in the main stream of a vigorous if super-
ficial intercourse between the Black Sea and the Byzantine
world on the one hand and Scandinavia on the other. But
in earlier times the hegemony of the Baltic lay rather with the
ereat Danish islands—Zealand, Fyen, Laaland—which bar
its outlet towards the Kattegat and the North Sea. These
—indeed, all the major Baltic islands—were at the same time
sufficiently large, sufficiently secluded, and yet sufficiently
accessible to support effective groups of seafarers and their
1T. Arne in Oldtiden, VII, 208.
2 Summary of views in H. Shetelig and H. Falk, Scandinavian Archaeology
(Oxford, 1937), chap. XIII.
28 ROME BEYOND THE IMPERIAL FRONTIERS

families under conditions of security hard to obtain on the

mainland. Through-shipping must often have sought them,
whether from choice or from adverse winds, and have con-
tributed to their wealth. As bases for trade or piracy they
had everything to offer. And their amenities extended to
the sounds and firths of the Jutish peninsula, where was also
a natural store of exportable amber. It is here then, at the
western end of the inland sea, on island and mainland, that
our maps show an early and continuing bias towards im-
orted wares. For example, of more than 500 Roman
vessels of bronze and 300 of glass known from Scandinavia, a
great majority come from Denmark.
Land- and sea-routes, sources in the south-east and in
the south-west, alike contributed to this concentration. At
one time or another in the first four centuries A.D. Denmark
drew its exotic wealth down the Oder and the Elbe from
central and southern Europe, and by the North Sea coast
from Gaul. The first check came with the great tribal
movements which broke the Rhine-Danube frontier of the
Empire in A.D. 258-60. Coin-evidence suggests that from
that time the Elbe route, hitherto fostered by the Thuringians
or Hermunduri beside its middle reaches (p. 22), ceased to
play any important part in the business. And though goods
from the Roman world continued in one way and another
to reach Denmark throughout the 4th century, the main
axis thereafter swung eastwards to the Swedish islands and
the Pomeranian-East Prussian seaboard, which has been well
described as one long quay at this period, busiest at its eastern
end about the delta of the Vistula.
The cultural unity and uniform wealth of Jutland and the
Danish islands in the first three centuries a.D. is difficult to
understand save in terms of some sort of political unity on
which history is silent. A West Baltic kingdom with a
sphere of influence extending to the mainland on the west
and south would explain the attractiveness of this region to
North Sea and continental traffic, and would make sense of
our distribution-maps (e.g. fig. 8). It is tempting to suppose
 FREE GERMANY: ROUTES AND MARKETS 29

that something of the old spirit of the Cimbri, who had

ravaged western Europe at the end of the and century B.C.,
survived here in the neighbourhood of their original peninsu-
lar home, and possibly derived a new force from the com-
pression imposed by the Imperial frontier. On the other
hand, a closer analysis of the Danish evidence has suggested
that this inferred West Baltic kingdom may have been
centred primarily, not on the Jutish or Cimbric peninsula,
but rather on the neighbouring islands, and that it spread
thence to the mainland at a relatively late date. About
A.D. 200 a break has been observed in the native culture of
Jutland.1 Pottery-types and brooch-forms change about that
time, villages are abandoned, cemeteries discontinued. New
cultural influences now come in from the islands, where the
tradition remains unbroken. The evidence requires further
thought, but meanwhile the picture is that of an island-
kingdom of the 2nd century a.p., with headquarters perhaps
on Zealand, abruptly extending its dominion into the
peninsula early in the following century and somewhat
drastically remodelling the way of life there. The 3rd
century, be it added, was the heyday of the vast peat-bog
deposits (p. 54), which, whatever their precise cause, sufli-
ciently show that great things were afoot at that time.
Then in the eastern Baltic we seem to have in the time of
the Later Empire a successor state or federation to this
western kingdom at a time when Goths from southern
Sweden and Burgundians from Bornholm had started upon
their continental ventures, incidentally unifying the east
Baltic coastlands and islands, and transforming them in
effect into a sort of “Gothic base’. Any more precise defini-
tion of the political situation in the two Baltic zones would
exceed the evidence, but such at least is the trend of the
archaeological distributions.
Norway and Sweden. Save for the Swedish islands (in
particular Gotland), Norway and Sweden may be described
as the ‘poor relations’ of Europe in the 1st and 2nd centuries
1H. Norling-Christensen in Acta Archaeologica, XIV (1943), 138.
30 ROME BEYOND THE IMPERIAL FRONTIERS

A.D. In contrast to Denmark, their distribution maps are

significantly thin until the great continental disturbances of
the 3rd and 4th centuries gave a new value to these peri-
pheral regions as refuges and bases. Not a single silver cup
of the type well represented in Denmark and northern
Germany found its way across the Baltic. Of the ‘chieftains’
burials’ which have provided a rich harvest of Roman
objects on the mainland, only one poverty-stricken example
is known from the Scandinavian peninsula. Even the
almost universal terra sigillata is absent there, except for a
single example from Gotland (fig. 13). A trickle of Roman
goods penetrated to the Oslo, Stavanger and Trondheim
fiords and to the lowlands behind Stockholm, and bronze
vessels of the plainest and cheapest type occur there (fig. 9).
But only when cut-glass vessels or rilled bronze bowls of the
Later Empire, or glass drinking-horns of the 4th-sth
centuries, began to circulate did the fringe of Norway—in
so far as a resistant geography permitted—attract imports of
the better class. The principal interest of these late dis-
tributions is that they reflect a new coastwise liveliness and
herald the era of the great North Sea folk-wanderings.
Iv CIRCUMSTANCES OF DISCOVERY
RAVES, settlement-sites, peat-bogs, undefined depos-
its have all contributed to the very considerable bulk
of Roman material now available from Free
Germany. In his recent lists, which exclude coins, brooches
and beads, Eggers: has catalogued no fewer than 2,257
find-spots, some of which have produced a considerable
array of objects. Nevertheless, the limitations of this
material are worth a moment’s thought at the outset, for its
recovery has been subject to many distorting conditions.
Thus, Denmark and the Netherlands have been more
systematically explored than most other regions, and, in
particular, the mass of find-spots along the Dutch coast is
due largely to this cause. Large areas, on the other hand,
such as Poland, which should produce much, have been
underworked. Outside Holland, few native settlements of
the time of the Empire have been excavated beyond the limes
or frontier. Graves, though better-recorded, are very liable
to be arbitrary in their representation. Rich burials such as
the so-called ‘chieftains’ graves’ probably overweight the
areas in which they happen to occur. On the other hand
the cremation-burials which represent the normal German
rite generally contain few grave-goods. Burial customs in
one region may specialize in the deposition of coins, in
another that of bronze situlae, without any necessarily exact
relationship to the respective distributions of those objects
amongst the contemporary living. Even in a rich cremation-
burial, associated objects may have become unrecognizable
through burning. Peat-bog dedications, again, are restricted
to the distribution of ancient marshes or meres; moreover
their discovery is necessarily more fortuitous than that of
burials which, even if not marked by mounds, commonly
occur in groups or cemeteries. Finally, in assessing the
1 See Bibliography, p. 182.
32 ROME BEYOND THE IMPERIAL FRONTIERS

evidence after it has been marshalled, how are we to interpret

differences of burial custom?—as indices of social stratifica-
tion? of chronology? of tribal distribution? Does a con-
tinuous, linear distribution reflect a trade-route, a campaign,
a migration? Or is it simply a natural zone of occupation?
And what, in any particular instance, is the significance of a
blank space on the map? Is it merely an accidental gap in
knowledge? If not, what is its probable meaning? These
are some of the questions which may fairly be asked—and
must often enough remain unanswered.
Burials are by far the most productive source of evidence,
and amongst them priority must be given to a remarkable
group of inhumations which, in spite of an approximate
uniformity, are widely spread in place and time.
(a) Chieftains’ graves of the Liibsow group, so named from
their richness and from the occurrence of five of them at
Liibsow, between Stettin and Kolberg in Pomerania. They
extend from the upper waters of the Oder in the south-east
across the German Plain to Fyen and Danish Slesvig in the
north-west, with a solitary poor relation in Norway, at
Storedal near Oslo (fig. 3). Their limit towards the west is
marked by two examples between the Elbe and the Weser.
Of the 32 burials recorded, all save two or three are inhuma-
tions. The graves are sometimes covered by a mound, but
are generally devoid of surviving surface-indication, and the
dead were often, perhaps always, placed in wooden coffins.
The wealth of their equipment, and the exotic character of
much of it, relegate them to a class by themselves. Roman
imports include banqueting services consisting of cups of
silver or glass, bronze wine-buckets, dishes, jugs, ladles,
strainers and ‘saucepans’, together with toilet objects such as
mirrors and combs; whilst among native products are silver
or bronze brooches, ornamented hair-pins, gold finger-rings,
silver and bronze buckles and other belt-fittings, gold-filigree
beads, bronze knives and shears, dice and gaming-pieces,
drinking-horns fitted with silver and bronze, casket-fittings,
1 Eggers, p. 50.
 CIRCUMSTANCES OF DISCOVERY 33

pottery, and sometimes spurs, but, be it noted, never

weapons. Joints of meat were also placed in the graves,

a en
“500 Metres
| “CHIEFTAINS” GRAVES
oot OF THE
: LUBSOW GROUP

« /NHUMATION
@® CREMATION

Scale: 2] 100 200 300 400 500 kK;/ometres

Fig. 3 (After H. J. Eggers)

together with a fermented drink made—as traces in the wine-
vessels show—from barley and local berries. (‘For drink’, says
34 ROME BEYOND THE IMPERIAL FRONTIERS

Tacitus, Germ. 23, ‘they extract a juice from barley or grain.’)

Men and women were buried with equal pomp, as befits
the honourable status ascribed to German women by Tacitus.
Inhumation was not the German Iron Age custom, and its
prevalence in the chieftain-burials has been ascribed, like
many of their contents, to the new contact with Roman
culture. The explanation is invalid, since at the time of a
majority of these burials, the normal Roman custom was
cremation. We are left with a puzzling paradox. For
1,000 years cremation had been the rule in Germany and
Scandinavia. Only when we reach back beyond a rather
shabby and reluctant Early Iron Age to the glorious Scandin-
avian Early Bronze Age do we find northern rites comparable
with those of the Liibsow group: the dead buried unburnt
with their cherished possessions about them—armlets of
gold, swords, daggers, axes, brooches, skins and textiles,
cups and bowls, even sprigs of blossom. It was as though
now, with the advent of new resources under the Early
Empire, the ‘new rich’ were reviving the magnificence of
their own remote past. There is nothing to suggest the
arrival of a new aristocracy from without; no known tribal
or national boundary defines the revolutionary mode. An
inter-tribal fashion, based upon an access of wealth, had
swept across central Europe in front of the organized and
masterful approach of the culture of the Mediterranean. It
represented the prerogative of an aristocracy or plutocracy
(merchant princes of the amber trade?) with ideas of its
own—a naive desire to continue intact in their graves
the delights of a pre-eminently peaceful and sumptuous
existence. That the divergent burial custom, like the in-
trusive riches, characterized a limited class is in accordance
with analogies from the ancient and modern world. In
India, for example, certain Hindu ruling families have long
inhumed their dead amidst subjects who adhere to cremation.
So about the chieftains’ graves of the Liibsow group the
ordinary tribesman normally continued to burn his dead,
humbly enough, in the traditional manner.
 CIRCUMSTANCES OF DISCOVERY 35

This sudden appearance of classical furniture in Free

Germany is remarkable, but not without parallel. Much the
same sort of thing had indeed happened not very long before
within the Celtic fringe some 300 miles further south. In
the sth and 4th centuries B.c. the opening up of eastern Gaul
to Mediterranean trade from Massilia (Marseilles) and
northern Italy had led to an influx of classical luxuries,
particularly wine and table-gear, into the region dominated
by wealthy Celtic aristocracies between the Rhine and the
Moselle. There had ensued in these parts new standards of
good living in what passed for the civilized Mediterranean
mode, with all the prestige of ‘Etruscan’ wine-jars, Greek
cups, and bizarre imitations of them. No cultural connec-
tion need or can be sought between these Celts and our
Liibsow Germans; the common factor is the response of a
flamboyant barbarism to sudden opportunity for exotic
display and affectation.
By what route or routes the Roman imports reached these
Liibsow chieftains is clear enough. Some contact from an
early date with the romanized West either by land or by the
North Sea cannot be precluded, but there can be no doubt
that the principal approach was from Italy direct, by the
Danube and the March to the systems of the Elbe and
the Oder. The map-distribution and the character of the
imports combine in proof, and a third factor coincides.
With scarcely an exception, the imports to which a southern
(mostly Italian) origin may be ascribed antedate the great
war against the Marcomanni and the Quadi of Bohemia in
A.D. 167-73 (above, p. 21). This war, and the prolonged
disturbances of which it was a peak, must seriously have
restricted trans-Alpine traffic; and it can be no mere chance
that it was only then and thereafter that the Liibsow ‘chief-
tains’ or their successors turned mainly to western markets for
their goods, their old sources of supply now largely closed to
them.
At the same time, a source of supply other than normal long-
distance trade may in some cases be suspected. Diplomatic
36 ROME BEYOND THE IMPERIAL FRONTIERS

gifts of plate to ingratiate native princes are attested by

Tacitus for Germany (above, p. 10) and by the Periplus for
the East (below, p. 116), and some of the more costly collec-
tions of classical table-ware from Free Germany may have
originated in this practice. The nearest approach to direct
evidence for it is supplied by the famous find from Hoby,
which constitutes the most remarkable assemblage of
classical metal-wares from any single spot in Free Germany,
with the exception of the Hildesheim hoard and the peat-bog
deposits.
In 1920 at Hoby, on the south coast of the Danish island
of Laaland, a richly furnished burial was found by chance and
excavated without method: The general features are,
however, clear. In an unmarked grave lay the skeleton of a
middle-aged man with his head towards the north-east. No
evidence of a coffin was observed, but on analogy it is not
unlikely that one was present. Beside the dead had been
buried two joints of pork and a magnificent table-service,
mostly in the more northerly part of the grave near the head
and chest of the skeleton. The equipment included a pair of
decorated classical silver cups placed on a bronze tray which
had been tinned internally; a plain silver cup or ladle to
which a vigorous zoomorphic handle of bronze had been
fitted by a non-classical craftsman; also of bronze, a situla or
bucket, a patera or saucepan, a jug; two bronze-mounted
drinking-horns; seven brooches of bronze or silver, one with
inlaid gold; two gold finger-rings; a bronze belt-buckle; a
bronze knife; a bone pin; and three clay pots. The datable
objects are of the end of the 1st century B.c. or the beginning
of the Ist century A.D., the time of Augustus. The burial
thus constitutes probably the earliest of the known chief
tains’ graves of the Imperial period.
The two cups at the head of the list are cylindrical two-
handled scyphi (a handle of one of them missing), and are
finely decorated in relief in a style inherited by the academic
age of Augustus from Hellenistic sculpture or painting of
1K. Friis Johansen, “Hoby-fundet’, in Nordiska Fortidsminder, Ul, Hefte 3.
 Ill

Roman bronze vessels and glass, with drinking-horns, etc., in a “chieftain’s’

grave at Juellinge, Laaland. (Two portions of the same grave.)
(See p. 41)
Roman vessels of bronze (1 and 4) and glass (2 and 3) from ‘chieftain’s’ graves
at Juellinge, Laaland. 1 and 4, 4; 2 and 3, 4 (See p. 41)
 iXs Silver cup, partially gilt, from the ‘chieftain’s’ grave, Liibsow I. bole

(See p. 43)

B. Glass cup with silver overlay from Varpelev, Zealand. co|bo

(See p. 71)
 (zs -d aag) (cS daag) § “yreurusd ‘pavrsotolqa J,
“puryoF] ‘osurzq wosyz “oydnf{Jo ajonjvys ozuOIg “a Wory ‘sIkY JO seNIwIs oZUOIG “V
i! g
VI
 CIRCUMSTANCES OF DISCOVERY 37
| the 4th century B.c. One represents the visit of Priam
to Achilles as recounted in the Iliad (pl. Ia, p. 20); the
) other derives its subjects from the legend of Philoctetes
and is probably based upon the lost Philoctetes of Euripides.
In both, certain details of the costume and ornament are
gilded. Both are signed in stippled letters by a maker
named Cheirisophos, the one in Greek lettering, the other in
Roman, though with the Greek form of the name. Cheiris-
ophos was evidently a Greek craftsman, working possibly in
Asia Minor, where Pliny places a centre of production, or
perhaps rather in Alexandria, with an eye to the Roman
market. The Philoctetes cup also bears, as these vessels
often do, an indication of the total weight of the pair, and
finally a graffito on the base of both gives Silius as the name
presumably of a former owner. Two comparable scyphi
are in the great treasure of Bosco Reale, near Pompeii, where
also the weights of important pieces and the names of owners
are inscribed. The bronze tray on which the Hoby cups
stood is of a kind more familiar to-day in the innumerable
red-glazed copies made by the Arretine potters. The jug is
of exceptional beauty, with a finely wrought frieze of tendril-
pattern and a charming little cupid at the base of a high and
graceful handle (pl. II, p. 21). The two-handled dish is of
comparable workmanship and has a scutcheon ornamented
in relief with a figure of Venus decked by cupids. The
situla has a floriate handle with swan’s-head terminals and
attachments adorned with cupid-masks. The type has a
wide south-east to north-west distribution across central
Europe, and even extends occasionally into Norway and
Sweden. Its Campanian origin is not in doubt, and its
bearers probably brought it via Aquileia and the Danube,
though a coastal route from the mouth of the Rhine is
possible. To a similar origin in southern Italy may be
ascribed the patera, which bears the maker’s stamp of Cn.
Trebellius Romanus. For the rest, the contents of the grave
are of non-Roman workmanship. The pair of drinking-
horns have knobbed or ‘baluster’ terminals of the longer
D
38 ROME BEYOND THE IMPERIAL FRONTIERS

and earlier type. Most of the brooches belong to a specific-

ally Danish series of the early Roman period; and, lastly,
the pottery is, as might be expected, also of local provenance.
In grave-deposits, cups and drinking-horns are found,
hospitably enough, in pairs according to an established
Roman habit, though the concurrence of the two types of
vessel—whether for different beverages or customs, or
whether the foreign silver was merely an added luxury—
is not explained. At any rate, Campania and Asia Minor or
Alexandria had contributed lavishly to the gear of the dead
man of Hoby, lying there with his clothing pinned about
him by a fine array of native brooches. The question
remains: How had a complete Roman drinking-service,
consisting of no less than eight pieces, come to this far-off
Danish island and to a single grave? By trade, it may be;
but here if anywhere we have an example of the diplomatic
presentation of gold and silver plate referred to by Tacitus.
For it has been plausibly suggested that the ‘Silius’ who had
owned the silver cup was none other than the C. Silius who
was legate of Upper Germany from A.D. 14 to 21.
Another burial of the same general kind but a century
later in date shows interesting variations in detail and is better
recorded. At Dollerup, about 10 miles west of Kolding in

1 Represented, for example, in the hoards from Bosco Reale and the House
of Menander at Pompeii, and, far away in pre-Claudian Britain, in the two
imported silver cups at Welwyn, Hertfordshire. Chemical analysis has shown
that, of a pair of native drinking-horns of early Imperial date found in a peat-
bog at Skudstrup, Kr. Hadersleben, in Danish Slesvig, one had contained mead
from honey, the other beer from emmer-wheat. Whether the pairs of classical
cups had similarly contained different kinds of wine or were merely a symbol of
hospitality is not known. The latter explanation seems unlikely. Whatever
the meaning of the custom, it was represented in no less than 17 of the ‘chief-
tains’ graves’ by pairs of metal or glass drinking-vessels, and had been taken
over from the Roman world with the goods themselves. When the source of
supply shifted from Italy to Gaul in the latter half of the 2nd century, cups were
no longer distributed in pairs; nevertheless, their German recipients con-
tinued to observe the Italian custom and made up pairs of odd vessels. Sce J.
Werner, ‘R6mische Trinkgefisse in germanischen Grabern der Kaiserzeit’, in
Ur- und Friihgeschichte als historische Wissenschaft (E. Wahle Festschrift), ed. by
H. Kirchner (Heidelberg, 1950), pp. 168-76.
 CIRCUMSTANCES OF DISCOVERY 39
| the Ribe district at the base of the Jutland peninsula, the
} discovery of pottery and other relics of various dates within
} the Roman period led in 1947 to a methodical exploration
which revealed a remarkable double burial.1_ Two graves lay
parallel with each other, roughly east and west, with less
than 1 m. between them, and the symmetrical but over-
lapping disposition of pots upon this intervening balk seemed
to indicate simultaneous burial. The bodies had been
placed in oak coffins of semicircular section, preserving the
contour of the tree, and had been laid out on, or shrouded in,
cow-hide. The pattern of the grave-goods showed that the
heads had been towards the west, although the actual bones
had disappeared save for some of the teeth. The more
northerly grave contained a large pottery vase near the head;
an iron brooch with silver inlay and traces of cloth where
the chest had been, recalling the words of Tacitus (Germ. 17)
that ‘the universal dress is the short cloak, fastened with a
brooch or, failing that, a thorn’; two drinking-horns with
bronze mouth-rings close to the face of the dead and, at the
pointed ends, bronze caps in the form of ox-heads; adjoining
these, fragments of silver hooks and bronze fittings, including
a buckle; a plain gold finger-ring; two rows of ornamented
silver studs; and at the foot two clay vessels—a cup and a
dish—with an iron knife and indeterminate objects of iron
and wood. The more southerly grave included similarly a
large pottery vase near the head; a silver brooch at the base
of the throat, with remains of cloth, hide, cord and a bead or
pendant of silver wire; two plain gold finger-rings where the
right hand had presumably been; towards the foot of the
coffin, two prick-spurs of bronze with silver filigree and iron
points, and two sets of three round silver filigree discs with
pierced lugs for attachment, traces of silver ornaments and a
blue glass bead; and, beyond these, an iron knife with remains
of a silver-mounted sheath, two bronze situlae or buckets
which may have been wrapped in cloth, two clay pots, and
1O, Voss and M. Orsnes-Christensen in Acta Archaeologica, XIX (Copen-
hagen, 1948), 200ff.
40 ROME BEYOND THE IMPERIAL FRONTIERS

two silver cups which had stood upon a round wooden tray;
and, lastly, fragments of a bone comb. The southern
grave had evidently been that of a man, but whether the
northern had been that of a woman is less certain.
Typical analysis of this varied assortment of objects,
particularly of the situlae, brooches, spurs and comb, indicates
a date not earlier than the end of the rst and more probably
within the 2nd century a.p. for the two burials. In the
present context interest attaches primarily to the bronze
buckets and the silver cups. The former, 25-5 cm. high, are
of double-conical form (Eggers types 24-9),! each with three
pelta-shaped feet, a pail-handle, and soldered handle-
attachments bearing human masks flanked by fish-heads.
Similar handle-attachments were found long ago in a mound
at Hgjsted (Rumperup) in the Holbaek district of Denmark
with a patera stamped by L. Ansius Epaphroditus, who
worked in Campania in the latter half of the 1st century
A.D. On the other hand, none of these double-conical
situlae seems to be recorded from Pompeii, so that their
vogue would appear to have been after a.D. 79. An example
from the Bosco Reale treasure, lost at that time, is typologic-
ally earlier; and a central date about a.D. 100, with wide
margins, may be ascribed to the Dollerup vessels, always
with the possibility of a long life prior to burial.
The two-handled silver cups found beside them are a more
difficult problem (pl. Is, p. 20). They are 7-9 cm. high,
have bands of zigzag, herring-bone and other simple linear
ornamentation round the upper part of the bowl with a
fringe of pendant triangles, and the disc-foot of the cup is
decorated with a further band of triangles. The handles are
flanked at the top (on the ‘thumb-plate’) by swans’ heads and
terminate below in a conventionalized animal head with
bulbous muzzle. All the ornamentation, including the top
surface of the handles, is gold-plated. The cups are of a
familiar classical form of the Ist centuries B.c. and a.D.,
but details point to provincial workmanship. The crude
1 See Bibliography, p. 182.
 CIRCUMSTANCES OF DISCOVERY 4I

ornamentation both on the thumb-plates and on the body is

not of the normal Mediterranean standard, and the free use of
emphatic triangular or denticular motifs is indeed a feature
of the local pottery from the same burial. Non-Roman
features have been observed in other cups of the series (p. 36).
Certainly the animal head at the base of the Dollerup handles
has nothing to do with Roman art; it is comparable with the
terminal heads of armlets, finger-rings and belt-fittings over
a wide zone of central Europe, extending from Jutland
through north Germany to Poland and lower Austria, with a
possible origin in the province of Noricum. Thus rather by
a process of elimination than by compelling evidence the
Dollerup cups have themselves been attributed to a factory
somewhere in that gold-producing Roman province. At
least it can be affirmed that, however ‘barbarian’ some of their
characters may be, they are as unlikely at this period to have
been a product of Free Germany as of metropolitan Italy. It
is sufficient to recall the words of Tacitus: “Heaven has
denied them [tne Germans] gold and silver—shall I say in
mercy or in wrath?’
Yet a third group of burials, somewhat later in date (perhaps
A.D. 150-200), may be described briefly to illustrate the range
of these rich deposits. Again in the island of Laaland, at
Juellinge, four graves were found in 1909, and three of them
were excavated with care.1 They were oriented north-
south, and the dead—all apparently women, young and aged
—had been buried in coffins, with the heads towards the
north. The graves were sealed by a number of large stones,
and with the coffins or in the filling above them were joints
of lamb; in one case, the complete forepart of a lamb at
one end and the hind-quarters at the other. The best-
preserved of the burials, now in the National Museum at
Copenhagen, may represent the series (pls. III, p. 36 and IV,
pp. 36-7). It was that of a woman who had died in her
thirties and had, incidentally, been lamed by a deformity of
her right leg. She lay on her right side with legs slightly
1 Sophus Muller, ‘Juellinge-fundet’, in Nordiske Fortidsminder, Ul, Hefte 1.
42, ROME BEYOND THE IMPERIAL FRONTIERS

flexed, and had been buried with her ornaments and clothing
about her: two gold-headed silver pins in her hair, an S-
shaped silver clasp with a gold pendant and two gold beads
at her neck, four silver brooches (two of them much worn)
on the chest and shoulders, two beads of amber and glass
under the right wrist, and a plain gold ring on the ring-finger
of the right hand. A mass of material under and about the
head probably represented a cushion. And held in the right
hand was a bronze wine-strainer in the form of a small bowl
pierced as a sieve at the end of a long and slender handle.
Beyond the head were the remains of a wooden box with a
bronze lock. It had been buried with the lid open, and
contained toilet implements: a bone comb, bronze shears, a
bone pin, and an S-shaped bronze knife. Beside the box was
a pair of Roman cut-glass beakers, and beside them again the
remains of a pair of drinking-horns with mounts of partially
silvered bronze. Their knobbed terminals are of the shorter
and later variety (compare above, p. 37). Beyond the horns
again was a bronze cauldron of a plain, convex-sided classical
type (Eggers type 40) relatively common in Free Germany in
the 2nd and possibly 3rd centuries a.p. Although it is
found also, as here, in wealthy graves, it was essentially the
‘poor man’s’ type, and as such had a wide vogue in the
poorer lands of the north, Sweden and Norway. Within
the cauldron, but formerly, it seems, on a wooden tray
placed over it, lay a bronze ladle into which the strainer
held by the dead fits; the two instruments were used together
in the process of ladling the drink from the cauldron into the
beakers or horns. A deposit in the bottom of the cauldron
showed on analysis that this had contained a fermented drink
made from barley and fruits of the countryside, and the soot
which covered the exterior of the cauldron was presumably
the result of the mulling of its contents.
The gentility implied by these Juellinge burial-groups with
their classical imports of bronze and glass must be tempered
by the observation that it was at best of a somewhat shabby
kind. At any rate, by the time that they were buried, the
 CIRCUMSTANCES OF DISCOVERY 43

imported bronze vessels were much worn and patched, and

the fact that one of them, a patera, bears the stamp of L.
Ansius Diodorus, who worked probably at Capua both
before and after the destruction of Pompeii in A.D. 79, has
no close bearing upon the date of deposit. Further, at least
one of the glass beakers in the grave which has been described
was seemingly damaged and useless at the time of burial.
Indeed, the whole aspect of the contents of the graves was that
either of a decayed heritage from better times or of an
economical selection of worn-out commodities, at least in
respect of non-native products. All things considered, the
date of the series must be well on in the and century.
Finally, it is only fair to the “Liibsow group’ to make some
reference to the first of the burials found in 1908 at Liibsow
itself, particularly since, unlike those described above, this
may possibly have been a cremation. The discovery was, as
often, accidental and the precise facts are uncertain,! but it
would appear that the remains had been placed in a grave
perhaps some 6 ft. in length with paved floor, stone-built
sides and a large cover-stone, surmounted by a cairn possibly
5 or 6 ft. high. At one end of the grave were clay urns
containing ashes which may or may not have been human,
and the lower parts of the side-walls showed evidence of
fire, supposedly from contact with hot wood-ash brought
from the funeral pyre. But in spite of the seeming intrusion
of the normal German rite of cremation, the grave has all the
general aspect and equipment of an inhumation-grave of
‘chieftain’ type. Apart from two drinking-horns, four
brooches (one of gold and silver), two silver pins and
bronze belt-fittings, the contents included a dozen objects of
Ist-century date from the Roman world: two silver cups,
two glass bowls, a bronze wine-bucket, jug and dish, two
bronze ‘saucepans’ or ladles, a mirror-plate of white metal,
and bronze shears and tweezers. The list closely repeats
those already given, and details may mostly be omitted. It
will suffice to note that the two-handled silver cups had stood
1B. Pernice in Praehistorische Zeitschrift, TV (1912), 126ff.
44 ROME BEYOND THE IMPERIAL FRONTIERS

upon the bronze dish, as at Hoby; and that they are of a

simple and graceful pedestalled form with analogies at Bosco
Reale and Hildesheim, plain save for a narrow band of gilded
leaf-ornament below the rim (pl. Va, pp. 36-7). As often,
under the base of each cup is scratched its weight. The glass
bowls—characteristically a pair, like the silver cups and
drinking-horns—are of a familiar 1st-century type with
vertical ‘pillar mouldings’ round the sides, and may well
be of Campanian manufacture. The bronze bucket is a fine
example of the type with ‘face-attachments’ for the handle,
i.e. with the plate attaching the handle-loops to the rim in the
form of a human head: in this instance, a female head with a
palmette-ruff and flanking animals’ (dogs’?) heads, possibly
indicating Scylla. A closely similar bucket is included in
the Bosco Reale hoard prior to a.D. 79, and the type is
regarded therefore as a Campanian product. It penetrated
fairly freely by way of the middle Danube to the Elbe-Oder
river-system and so to Denmark, with two outliers in
Norway (one of them as far north as the Trondheim fiord)
and three in Sweden. The larger of the two ‘saucepans’ or
paterae is tinned internally and in part externally, and has an
eatly type of handle with a semicircular piercing flanked by
swans heads (fig. 7). This type is particularly common in
Bohemia, which may well have acted as an intermediary
towards the north. Like the rest of the imported bronze-
work in the grave, the bucket may be ascribed to Capuan
or at least to Campanian workshops.
These four examples, from Hoby, Dollerup, Juellinge and
Liibsow itself, will suffice to illustrate the remarkable
character of the ‘chieftains’ graves of Liibsow’. The
durability of many of their contents, with the patent fact that
some of these had obviously been long in use before burial,
complicates their chronology, but most of them are safely
included in the bracket a.p. 50-150. Their distribution,
from Jutland to Poland, is too diffuse to conform with any
conceivable political amalgam. At the same time, their
astonishing uniformity indicates something more than
 CIRCUMSTANCES OF DISCOVERY 45
chance. Their furniture shows us a well-to-do society
intertribal in scope and shaped by uniform conventions and
aspirations: the possession, above all, of a drinking-service,
with cups (and drinking-horns) correctly duplicated in the
classical tradition, and with containers, ladles and dishes of
the orthodox Italian kind. The wine, indeed, in these
northern parts had not always seen the Italian sun; the proud
owner of Campanian bronze or glass or Graeco-Roman
silver must often enough have prefixed his wassailing with
an apology for serving the humble mead and cranberry-
juice which had contented his ruder ancestors. Rome,
which had imposed a new culture upon the conquered west,
had simultaneously imposed a new snobbery upon the
unconguered north. These mute, inglorious ‘chieftains’ and
their wives, we may suppose, were not immune from this
human frailty. But they have richly contributed to archaeo-
logy if not to history.
(b) Other burials. It may be admitted that to call the
wealthy, placid Liibsow folk ‘chieftains’ is to conform with
an archaeological convention rather than to state an ascer-
tained or even probable fact. Nothing that we know of
Teutonic chieftainship would lead us to suppose that a
prince of Tacitus’s Germania would venture into the here-
after without at least a token armament. It may be that con-
tact with the Roman world had brought into being, during
the effective tenure of the neighbouring Roman limes, a
tolerably settled stratum of native ‘new rich’ not wholly nor
even largely identical with the impulsive military aristocracy
of the German tradition. If so, the seeming fact that these
‘new rich’ faded out when major disturbances began to alter
the pattern of Free Germany in the latter part of the 2nd
century is easy to understand. A hundred years later, their
place was taken in central Germany by other ‘chieftains’,
wealthy indeed, but for the rest of a more traditional
kind.
These later ‘chieftains’ have been studied particularly in the
region of the Saale and its tributaries south of Halle and
46 ROME BEYOND THE IMPERIAL FRONTIERS

Leipzig.t They were buried north-south in large coffins or

wood-lined graves, were mostly men—young to middle-
aged—dressed and spurred, with gold finger-rings and silver
brooches, Roman paterae, bowls and dishes of bronze, native
and Roman pottery, Roman glass, and with silver spear-
heads or arrowheads by their side. Roman coins and the
types of spur and brooch indicate dates from the end of the
3rd century to the middle of the 4th.
In these same regions are many other humbler graves, at
first cremations, later including inhumations. The sugges-
tion of the evidence as a whole is pretty clearly that of a
military aristocracy and their followers within raiding-
distance of the shrinking Roman frontiers at a time when
central Europe was seething with tribal movement and loot
was abundant.
For the rest, little need here be said of the large number
of burials, mostly cremations, which have yielded Roman
goods in other parts of Free Germany. The bulk of the
material of our study is derived from them, and their
geographical extent is remarkable. In the north they reach
the head of the Trondheim fiord in Norway and the province
of Vaasa in Finland. The Danish and Swedish islands and
eastern Jutland are full of them. Further south they clutter
the environs of the great river-valleys and occur more
sparsely between them. Frequently the settlements repre-
sented by these burials have not been identified; it may be
assumed that mostly they consisted of poor huts, such as
those described on p. 49.
A single grave-field of the 3rd century A.D. in Hanover will
here serve to illustrate in a general way the whole vast series.
At Helzendorf, Kr. Grafshaft Hoya, in that province were
found in 1936 a number of cremation-burials which produced
a medley of Roman and native craftsmanship,? the latter
1 'W. Schulz, ‘Die Skelettgraber der spitr6mischen Zeit in Mitteldeutsch-
land’, in Mannus-Bibliotek no. 22 (Leipzig, 1922); and Schulz, Leuna: ein
germanischer Bestattungsplatz der spdtrimischen Kaiserzeit (Deutsche Academie
der Wissenschaften zu Berlin, 1953).
2 Germania, XXIII (1939), 168ff.
 CIRCUMSTANCES OF DISCOVERY 47
usefully dated by the former. Most of the graves consisted
of a hole in the ground in which a bundle of burnt bones had
been placed in a cloth wrapping. A few, however, of
larger size (75-100 cm. wide and up to 85 cm. deep) con-
tained Roman bronze vessels with their bases set firmly in
the underlying sand. Three of these were simple buckets
of the so-called ‘Hemmoor’ type, a Gaulish or Rhineland
product characteristic of the 3rd century (p. 78). In two of
them with the ashes were remains of a bone comb, and in
one the oxidized metal had preserved traces of a bedding of
leaves beneath the bones. Another grave produced a
bronze bucket or situla with cylindrical neck above a carin-
ated body and with loop-attachments in the form of a
woman's head between the fore-parts of fish. This is a late
variant of the ‘face-situla’, which occurs freely from Bohemia
to Zealand—an exception to the rule that after the middle of
the 1st century Bohemia withdraws into the background as a
Roman market. Presumably the convenience or attractive-
ness of these situlae as ash-containers led to a circumvention of
the normal difficulties of trade. In the Helzendorf example
the ashes had been bedded in straw. Yet another of the
graves contained a bronze tray with vertical sides and vine-
leaf handle-loops, the loops or hooks themselves terminating
in panthers’ heads. Within the tray were a bone comb and
clear traces of the linen bag which had contained the burnt
bones, and in the same grave was a hand-made native pot.
Other objects from the cemetery included a number of native
pots, a gold finger-ring with lapis lazuli intaglio, and a terra
sigillata beaker of 3rd-century type with barbotine decoration
(Ludowici form VMeg), probably of Rhenish manufacture.
Within the purlieus of the Roman frontier the native
villagers were liable on rare occasions in the 3rd or 4th
century to adopt the Late Roman custom of inhumation.
Thus at Laisack, north of the upper Danube (Raetian)
frontier near Neuburg, was found a flat grave containing a
skeleton with three bronze arrow- or lance-heads, a silver
buckle and four pots, of which one was a black-glazed
48 ROME BEYOND THE IMPERIAL FRONTIERS

Rhenish beaker of familiar late type, inscribed [RE] PLE

(‘fill up’) in white slip.1 This beaker may have wandered
200 or 300 miles from its place of manufacture, but the mode
and route of its wanderings in this period of Roman with-
drawal and spasmodic German advance can only be guessed.
(c) Settlements. Any attempt to describe the homesteads or
villages of Free Germany would exceed the scope of this
book. We are not here concerned with the social and
economic life of the German or Scandinavian tribesmen save
in so far as is necessary to an understanding of the direct or
indirect impact of Roman material culture upon them.
The remains of their houses accord with the remarks of
Tacitus in that they show a simple architecture of timber
partially smeared with clay, without masonry or tiles, and are
isolated one from another perhaps partly to minimize the risk
of fire. Typical examples from Denmark? are about 50 ft.
long and 15 ft. broad, with lines of wall-posts supplemented
externally by walls of peat or turves, and seemingly with
turf-covered or thatched roofs, the construction of which was
deliberately simplified by the moderateness of their span.
The entrance is in the middle of one of the long sides, and the
hearth, of stones covered with clay, is more or less central,
but not immediately opposite the door. A stone mortar may
lie on the clay floor, and there is a tendency for sherds to
accumulate round the margins. Here and there scraps of
Roman pottery or metalwork, or a Roman Republican or
Imperial coin, occasionally of gold but more usually of silver,
may be found, but as preserved the imported material from
these simple homesteads is not comparable in quantity or
quality with that in the inhumation graves further east. These
long houses were intended to accommodate cattle as well as
the farmer and his family, the living quarters being commonly
up-wind, at the eastern end.
Sometimes the distinctive long-house plan is replaced
1 Germania XVIII (1934), 117ff.
? For these, seeJ.Brondsted, Danmarks Oldtid (Copenhagen, 1940), Ill, 107ff.
and 243ff.
 CIRCUMSTANCES OF DISCOVERY 49
by a smaller type, more nearly square and with only two or
three posts on each side. The entrance is still central, and the
hearth or hearths are on the major axis. Crofts of this kind
were indeed probably normal throughout Free Germany.
For example, at Wittislingen in Bavaria, on the banks of the
Egau, a southern tributary of the Upper Danube in the old
province of Raetia, a Roman building of stone was
destroyed (as finds show) during the great invasion of the
Alamanni in A.D. 259-60.!. Sometime later, but within the
period of the Roman Empire, German settlers occupied
the site and built a group of small huts, each about 12 ft.
square with floor sunk 1} ft. below the surface and with
three wall-posts on two opposite sides. The sinking of the
floor is incidentally a German habit repeated in Saxon huts
both in Holland and in England.
Along the Frisian coast between the mouths of the Rhine
and the Elbe, local conditions imposed variations in the
layout of the homestead. The land surface hereabouts was
liable to flooding, and many of the huts and byres were
therefore built on extensive artificial mounds or fterps heaped
up to a height of some 10-20 ft. on the local sandy clay or
alluvium. These mounds have not infrequently been
absorbed to a great extent by the aggradation of the sur-
rounding plain, but attention has been directed to them
from time to time in the Dutch sector by intensive agri-
culture, and these discoveries have been skilfully exploited
and supplemented by a society formed for the purpose under
the leadership of Professor A. E. van Giffen.? Unfortunately,
any distribution-map founded on these researches is partially
falsified by the accident that the adjacent stretch in Germany,
between the Dutch frontier and the Elbe, where a similar
type of ancient occupation may be presumed, has been used
in modern times predominantly as pasturage and remains
undisturbed and unexplored.
1 Germania, 1952, pp. 287ff.
2See reports of the Vereeniging voor Terpenonderzoek from 1918 on-
wards; and Germania XX (1936), 4off.
§0 ROME BEYOND THE IMPERIAL FRONTIERS

The mounds range in date from the pre-Roman Iron Age

to the Carolingian period or later, though the known pre-
Roman examples are rare. A single notable example will
serve to represent the sequence. North-west of Groningen
lie large numbers of the mounds—variously known as Terp,
Warf, Werft, Wurt or Wierde—in the vicinity of existing or
former streams on the flat Frisian coastland. The summit
of the largest mound is crowned by the parish church of
Ezinge, at the head of the little town which covers the gentle
south-eastern slope. Towards the north, a considerable
sector of the mound was carefully excavated by Dr. van
Giffen in 1931-4, and revealed a succession of timber farm-
steads ranging in date from about 300 B.c. to A.D. 400 or later.
To-day the mound is some 1,500 ft. in diameter and rises at
the centre about 17 ft. above the natural soil. This height
is due partly to the accumulation of occupation-material
(remains of floors and turf-walls, posts and wattle, cattle-
dung, general debris) through the centuries, but owes much
to a periodical building up of the site by the deliberate
addition of clay or dung in order to maintain the super-
imposed homestead above a flood-level which was itself
rising with the gradual ageradation of the surrounding plain.
The earliest house (Phase VI) lay in fact on the old natural
surface; but it was succeeded by a platform which grew in
height and diameter through four subsequent phases of
occupation. By the Roman period (Phases II-III) it was
114 ft. high and about soo ft. in diameter.
With the exception of the crude square hut-pits which re-
present the ultimate occupation (Phase I) by intrusive Anglo-
Saxons about A.D. 400, the successive homesteads are based
throughout uponcertain uniform elements. These are of con-
siderable interest in the general history of the house-plan, but
cannot be more than referred to here. Briefly, they com-
prise oblong buildings commonly divided into a ‘nave’ and
two ‘aisles’ by internal lines of posts, three or more in
number on each side. The bays of the aisles are sometimes
partitioned into stalls, and may be fronted along the sides of
 CIRCUMSTANCES OF DISCOVERY $1
the nave by a low wickerwork rack to carry fodder for the
stalled cattle. Where the stalls are absent, a hearth (or two
hearths) in the nave indicate human occupation. Sometimes
both usages are combined in the same building: stalls and
racks towards one end and a hearth towards the other. The
whole indicates a simple farming economy of a familiar early
medieval type, here going well back into the pre-Roman
Iron Age. The buildings range in size from a small com-
bined house and byre 31X23 ft. to a monstrous barn with
continuous stalls and no hearth, upwards of 76 ft. long. The

aisles (stalls) are about 5 ft. broad. In some instances there

is a hint that the main posts may have curved inwards as
crucks. The door was placed as nearly as possible centrally
either in the end or in the side, or in both.
Each phase was marked by rebuildings which make it
difficult to establish a unitary group at any one moment.
The clearest ensemble is included in the homestead-plan
characterizing Phase V which, though approximately of the
3rd century B.c., must closely have resembled its successors
of Roman date. The main element in this ensemble is a
$2 ROME BEYOND THE IMPERIAL FRONTIERS

large byre of the kind described, with an aisled addition to

one side and flanked on the other by two smaller-aisled
dwellings with hearths, and a third combined byre and
dwelling. The whole makes a satisfactory reconstruction
(fig. 4).
It may be added that amongst the contents of the Roman
levels (Phases II-III) were bronze brooches of the 1st century
A.D., sherds of plain and decorated terra sigillata of the 2nd
century, and a bronze statuette of Jupiter. Roman statuettes
may indeed be described as characteristic of the region (pl. VI,
p- 37): no fewer than 34 of them, from 28 find-spots, have
been found in the provinces of Gréningen, Friesland and, in
a minor degree, Drenthe. It has indeed been suggested that
their numbers are such as to indicate some sort of ‘statuette
cult’ amongst the Frisian farmers, but if so the statuettes are
too various to indicate its trend.
In addition to the statuettes, it has been calculated that the
provinces mentioned have produced over 50 finds of terra
sigillata and 200 coins of the tst-4th centuries A.D., mostly
from terps. The calculation is probably an underestimate.
But it is scarcely necessary to emphasize that in the Frisian
culture there was nothing which can be described as Roman
in any significant sense. The Frisians, like the neighbouring
Batavians to the south-west, had aided the Romans during
the campaigns of Drusus and Germanicus, but revolted in
A.D. 25 in consequence of tax-oppression and again in A.D.
69-70, when, once more with the Batavians, they achieved
an enduring political independence. They appear, however,
to have been an essentially peaceful folk, primarily stock-
farmers (sheep and cattle, with some horses and pigs), but
with an itch for the sea. They traded along the coast and
perhaps via the canal whereby Drusus linked their territory
with the Rhine in 12 8.c.; acting as carriers to and from the
wealthy markets of Jutland and the Danish islands, and
doubtless bartering their livestock. In the 3rd century a.p.,
if not before, they were transporting cattle up the Rhine.t
1H. Wilkens in Hansische Geschichtsblatter, XIV (1908), 310.
 Vit

A. Silver parade-helmet from the moss of Thorsbjerg, Denmark. oe

(See p. $7)

B. Silver bow] imitating glasswork, from Haagerup, Fyen. 3

(See p. 70)
 VIII

vee
ASS
Aiea,
 CIRCUMSTANCES OF DISCOVERY 53

Enough has been said to indicate broadly the environment

in which Roman goods circulated, to a surprising extent,
throughout a great part of Free Germany. Save in so far as
Roman weapons and armour influenced the German war-
gear (and vice versa), the impact of the one culture upon the
other was, at any rate at first, of no great significance. The
two societies were basically far too disparate for fruitful
interaction, and Roman sherds or coins on the trampled
floor of a German hut meant no more than did the Arretine
dishes which strewed the squalid wigwams of Cunobelin at
Colchester. But this contact at least served to show the
resurgent German tribesmen of the Middle and Later Empire
in what directions lay the rewards of victory, and must to
that extent have stimulated and even oriented the earlier
phases of their wanderings. The alluring taste for half
understood exotic things has often enough helped to bend
the course of history.
(d) Loot. It cannot be doubted that in the turbulent times
of the Later Empire the chronic pillaging of the Roman
frontier-zone by Alamanni, Franks and other German tribes
or confederacies, to say nothing of free-lance piracy, con-
tributed to the diffusion of Roman goods in Free Germany.
In the next section the assemblage of Roman and romaniz-
ing material in the peat-bog hoards of Denmark will suggest
sources of this kind. Rarely, however, can we expect dis-
criminating evidence as between trade and plunder, and this
heading need not therefore detain us long. Two examples
will suffice. A bronze situla of the ‘Hemmoor’ type (p. 78)
and probably of 3rd-century date was recovered long ago
from a barrow at Bjorska in the Vastmanland province of
Sweden, some 60 miles N.W. of Stockholm. It had been
used ultimately to hold the ashes of the dead, but it bears the
inscription APOLLINI GRANNO DONVM AMMILLIVS CONSTANS
PRAEF TEMPLI IPSIVS VSLLM (‘Dedicated to Apollo Grannus by
Ammillius Constans, warden of the temple’) and, as temple
property, is very unlikely to have been given or traded into
native hands. The cult of Apollo Grannus had a special
B
54 ROME BEYOND THE IMPERIAL FRONTIERS

vogue in Raetia, and the distribution of the Hemmoor type

is consistent with an origin in that quarter. The region was
traversed by the Alamanni when they broke through the
limes between the Danube and the Rhine in a.p. 213 and
again in 260, and the situla doubtless found its way to its
remote find-spot as loot on some occasion of the kind.
Again, at a much earlier date, in wooded country some-
where between the Ems and the Weser had occurred one of
the greatest disasters that ever befell the arms of Rome:
the utter defeat and massacre of the three legions of Varus in
A.D. 9 by the tribesmen of Arminius. This cataclysm
marked the end of systematic Roman exploitation, political
and commercial, towards the Elbe. But meanwhile the
whole of the vast equipment of a considerable army passed
into German hands; and, though we cannot hope to identify
much of it at this long range, the remarkable concentration
of Augustan gold around the scene of the catastrophe can
scarcely be accidental. It is a fair guess to recognize here
some of the spilled treasury of Varus (fig. 5). On the other
hand, the famous silver dinner service found buried at
Hildesheim near Hanover at a considerable distance from the
nearest Roman settlement may in fact be the campaigning
outfit of a Roman commander, but cannot, as has been
suggested, be ascribed to Varus or Germanicus, at any rate
in its entirety, since, whatever the dates of its component
parts, one of the pieces bears what is almost certainly an
Antonine inscription.
(e) Peat-bog or moss deposits. Last but by no means least
are the great assemblages of equipment which have been
found during the past century in the mosses of Denmark.
These mosses represent former meres or inlets of the sea
which have in one way or another become isolated and so
have degenerated first into marshland and then into peat-
bog. Sealed successively by water and peat, objects often
of a perishable kind have been preserved with a completeness
rare in surface deposits, and the routine work of peat-cutters
1°W. Knapke in Acta Archaeologica, XIV (Copenhagen, 1943), $8.
 CIRCUMSTANCES OF DISCOVERY 55
has combined with the vigilance of the Danish authorities to
recover great quantities of this submerged material. Its
presence in the former meres requires explanation, although
any explanation must now in part be theoretical. Certain

Fig. 5 Distribution of Augustan gold coinage resulting possibly

from the Varus disaster, A.D. 9. (After W. Knapke)

it is that in Scandinavia from the Stone Age onwards it was

customary to dedicate objects of value—large flint axes, flint
daggers and saws, bronze axes and swords, bronze trumpets,
gold vessels, amber ornaments—by depositing them in the
56 ROME BEYOND THE IMPERIAL FRONTIERS

eround, in streams or springs, or in lakes. In the pre-Roman

Iron Age this custom was continued and, it seems, given a
new turn. Thus at Hjortspring in Danish Slesvig a 30-ft.
boat was sunk some time in the 4th or 3rd century B.c. with
fragments of ring-mail, 50 shields, eight swords, 170 spears
with iron or bone points, and animal carcasses, and it is
noteworthy that many of the weapons had been rendered
intentionally unserviceable before their deposition. This
procedure has recalled the statement of Orosius (V, 16) that,
after their victory over the Romans at Orange in 105 B.C.,
the Cimbri and Teutons mutilated the captured gear, threw
some of it into the river, drowned the horses and hanged the
captives. “All booty was destroyed in accordance with a
new and unaccustomed vow’—new and unaccustomed at
any rate to the historic world. Caesar (B.G. VI, 17) also
refers to a custom amongst the Gauls (a term which for him
included western Germans) of sacrificing ‘such living things
as they have captured’ and of gathering their other booty into
conspicuous heaps in hallowed spots. Similarly, at Toulouse,
as Strabo recounts (4. I. 13), in 106 B.c. a Roman consul
pillaged a great treasure which had been ‘stored away partly
in sacred enclosures, partly in sacred lakes’ by its Celtic
owners. And on the Celtic fringe, a great mass of Iron Age
weapons and other gear found in a bog in Anglesey in 1944
is best explained as the product of religious dedication. In
one way and another, there is ample precedent for a series of
rich and famous peat-bog hoards from Slesvig, Jutland and
Fyen dating from the troubled 3rd and 4th centuries a.p.
The contents of these peat-bogs need not be catalogued
here in great detail, but three of the discoveries may be
summarized to illustrate their range and character.1 One
of the earliest of them (1858) was that of Thorsbjerg, so-
named from the adjacent “Hill of Thor’, a dozen miles north-
east of the town of Schleswig, now in Germany, but until
1864 in Denmark. The site is that of a former mere which
may anciently have had an outlet to a neighbouring rivulet
1 See C. Engelhardt, Denmark in the Early Iron Age (London, 1866).
 CIRCUMSTANCES OF DISCOVERY 57
and so to the Firth of Sli, three miles to the south. The mere
had become a mass of peat 11 ft. thick, and it was in a limited
area within the lowest level of this peat, and beneath it, that a
remarkable collection of objects was brought to light. A
majority of them consisted of arms and armour, but orna-
ments, tools, agricultural implements and even clothing
were included. Exceptional were parts of two Roman
parade-helmets: one of thin bronze ornamented with wave-
or fire-patterns, a wreathed star and possibly a thunderbolt,
the other of silver lined with bronze and partially covered
externally with thin gold plates. The latter helmet (pl. VI,
p- 52) is composed of a face-piece (now lacking a vizor,
if it ever had one) and a strapwork crown hinged to it, and is
enriched with conventionalized hair and bands of small
bosses on the crown and fringing the face. It would be a
notable ‘find’ anywhere, but in this remote spot its interest
is enhanced by the mystery of its immediate origin—
scarcely an ordinary article of trade, but whether the product
of a raid upon the Roman frontier or a diplomatic gift from
the Roman world to some native chieftain can only be
guessed. Certainly many other objects of Roman proven-
ance were associated with it: fragmentary coats of mail;
circular wooden shields,: either with round bronze bosses of
Roman type—in one instance bearing the dotted name
AEL. AELIANVS—or with conical bosses of native type, one
with a Runic graffito; the bronze-fitted wooden scabbard
of a Roman short-sword, and the wood or metal hilts
and fittings of many other swords either of Roman manu-
facture or native adaptations; javelins (pila or angons);
brooches, buckles, sandles and other gear. Two discs of
gold- and silver-plated bronze incorporate or have been
augmented by animals of a semi-Oriental type which the
It may be observed that on Roman monuments, mostly earlier than this
hoard, Germans, like Celts, are shown ordinarily with oval or polygonal-
oblong shields, and the Hjortspring hoard (above p. 56) includes oblong
shields with rounded corners. Circular German-shields occur on the Column
of Marcus Aurelius, however, and the round shield was characteristic of the
Folk-wandering period. Cf. Tacitus, Germ. 43.
58 ROME BEYOND THE IMPERIAL FRONTIERS

Goths spread across Europe from their far-off kingdom by the

Black Sea. With all this was a great variety of more specific-
ally native material, though with recurrent romanizing
features: iron spears, spear-shafts of ash 8-10 ft. long; long-
bows of ash, about 5 ft. long, and iron-shod arrows; whet-
stones, horse-harness, parts of wagon- or chariot-wheels,
pottery; and five articles of woollen clothing, comprising
two cloaks woven in a twill-pattern and in one case green
with a yellow and black fringe, a shirt of which the long
sleeves are ornamented with a diamond-pattern in the cloth,
and two pairs of trousers with loops for the belt and ‘feet’
of a different and patterned cloth (pl. XIIIa, pp. 84-5). With
the great hoard were 37 Roman denarii ranging from Nero
to Septimius Severus (A.D. 193-211) and sufficiently indicating
the 3rd century as a terminal date for the deposit.
About 20 miles north of Thorsbjerg, within three-quarters
of a mile of the Als Sound of which it anciently formed an
inlet, lies a narrow valley, now floored with peat, the famous
moss of Nydam. Beneath the peat, at a depth of 4-7 ft.
and covering an area of 10,000 sq. ft., was found in and after
1858 perhaps the most famous of the Danish bog-hoards.1
Unfortunately, systematic investigation was interrupted by
the German invasion of 1864 and, in the words of the Danish
investigator, ‘the subsequent excavations at that spot, under-
taken by German Princes and by a Prussian Baron, do not
seem to have been carried on with the necessary care and
intelligence’. Nevertheless, a wealth of material had already
been recovered, representing much the same range as the
Thorsbjerg hoard, but with additions and clearer detail.
The outstanding feature of the Nydam hoard was the
presence of three boats, of which one, a clinker-built, sail-less
yacht 77 ft. long, was removed almost intact. With the
structure of these vessels we are not here concerned; it will
suffice to note that they had been pierced and deliberately
sunk, and that the objects found in and around them had
apparently been on board them at the time of the sinking.
1C, Engelhardt, Nydam Mosefund (Copenhagen, 1865).
 CIRCUMSTANCES OF DISCOVERY 59

These objects occurred in heaps tied together or wrapped in

linen, and included 100 swords, some of them stamped
with makers’ names in Roman lettering (rIcvs, RICCIM,
COCILLVs and others); between 500 and 600 heads of javelins
and spears; 40 bows and 170 iron arrowheads; a scythe-
blade and other tools; dress-fittings and ornaments; and
34 Roman denarii, extending from Vitellius to Macrinus
(A.D. 217-18). The date of deposition was presumably,
therefore, not distant from that of the Thorsbjerg hoard,
though the character of certain of the Nydam objects
suggests that they may be somewhat later, extending into the
4th century.
A third peat-bog hoard was recovered about the same time
from the Vimose or Moss of Wi (i.e. temple), near the Firth of
Odense in the north-east of the Danish island of Fyen.1
Of 2,000 objects—swords, one with the stamp Tasvit,
sheaths, spears, bows and arrows, a quiver, shields, and much
other gear together with wooden bowls and pottery—some
special mention may be made of two. One of these is a
bronze ornament of Roman workmanship in the form of
a griffin’s head, probably from a helmet. The other is a
complete shirt of mail, the only intact example from Free
Germany of a type of armour abundantly represented in the
bog-hoards by fragments. The shirt is of bronze-riveted
iron rings about 4 in. in diameter, has short sleeves and is
about 34 ft. long, extending to above the knees (pl. VIII, p.
53). Similar shirts of mail are familiar as worn by soldiery
on the columns of Trajan and Marcus Aurelius and other
Roman monuments, but the type is probably of Asiatic
origin and was known to the Germans long before the
Imperial period—witness the Hjortspring hoard mentioned
above. With the Vimose hoard were coins of Marcus
Aurelius and Faustina the Younger, implying a date after
A.D. 180, but insufficient for precision.
A review of these hoards and of others like them indicates
certain common and significant features. First, they are
1C, Engelhardt, Vimose Fundet (Copenhagen, 1869).
60 ROME BEYOND THE IMPERIAL FRONTIERS

mostly within easy reach of the sea, and are characteristic of

the wealthiest region of Free Germany—namely, the eastern
side of the Jutland peninsula and the Danish islands. Outside
this region the only find at all comparable is a later and much
simpler one from Kvalsund in Norway. Secondly, each
hoard certainly or probably represents a single deposit
rather than a gradual or casual accumulation. Witness
the bundled cargoes of Nydam, with shields in one part,
arrows and wooden vessels in another, a group of swords in
yet another; or the fact that at Thorsbjerg specific classes of
objects were likewise found in separate heaps—for example,
several layers of shields pinned together by a javelin which
had been thrust through them, or a concentration of gold
objects, or of mail, in which spear-heads and small arms
might be collected and wrapped. Thirdly, most of the
objects had been deliberately damaged, often not in a fashion
which can be ascribed to the incidents of battle. The shield-
bosses are bent and crumpled and broken, a wooden sheath
from Thorsbjerg is cut in two and the pieces placed one on
the other, many sheaths are deprived of some or all of their
mountings, mail shirts are cut and rolled up and the pieces
sometimes placed carefully in clay pots, lances are fantastic-
ally twisted. At least one of the Nydam boats had been
intentionally and drastically destroyed, and the two boats at
Kvalsund had been deliberately shattered so that no two
timbers were left fastened together. More than that, though
human skeletons are almost certainly absent from the bog-
finds, the bones of slaughtered animals, particularly horses,
are abundant, and the butchering has been carried out with
an astonishing thoroughness. Thus a horse’s skull from
Nydam bears no fewer than to sword-cuts, and others
show similar treatment. At the Vimose many horses’
skulls and bones were found splintered and hewn to pieces
beside chopping-blocks which had been extensively used.
Another horse at Nydam had an iron arrowhead in the
left shoulder and a second in the ribs. Detached heads of
horses at the same place still had their bits in their mouths.
 CIRCUMSTANCES OF DISCOVERY 61

Equipment and animals alike had been mutilated system-

atically before being consigned to the mere or inlet in which
the peat has subsequently covered and preserved them.
Two other observations may be added. The hoards, in so
far as they are datable, belong to the period of the Later
Empire, the 3rd or 4th century a.p. And they are linked
predominantly with the Roman world. Many of their
contents were made within the boundaries of the Empire;
many others betray the native craftsman with a knowledge
of Roman things; and, if from time to time a semi-Oriental
element, such as a Scythian-looking animal pattern or
perhaps the bows, may have been derived from south-eastern
Europe, we have to remember that by this time a ‘Gothic’
zone extended from the Baltic to the Black Sea and that
from this zone tribal groups—Burgundians, Alamanni and
others—were pushing westwards and familiarizing the West
with south Russian modes.
If now all these various facts are added up, they make a
not unimpressive total. At a number of points readily
accessible to a society based largely upon the sea, great
assemblages of war-gear were sacrificed in a traditional
manner, by slighting and submersion. The picture is, as
we have seen, exactly that painted by Orosius speaking of
the Cimbri and Teutons at Orange, or by Caesar speaking
of the Gauls in general. Whoever the rulers of the Cimbric
peninsula were in the 3rd century a.D., they or their tributary
tribesmen were bringing in to their religious centres great
quantities of loot, mostly obtained either by piracy in the
Baltic or in the English Channel, where the British and
Gaulish shores were now specifically fortified against Saxon
sea-raiders, or, perhaps more probably, from battlefields
somewhere within reasonable range of the Roman frontier.
In the absence of precise dates for the hoards, a closer context
is not feasible. Suffice it to recall that in a.p. 213 the con-
federation of German tribes known as the Alamanni burst
through the Raetian frontier and heralded a long age of
increasing German pressure on the West. In a.D. 257-8
62 ROME BEYOND THE IMPERIAL FRONTIERS

another confederation, the Franks, from between the Ems

and the Rhine, swept through Gaul and into Spain. Two
years later the Alamanni were again on the move; they
carried destruction across the limes into Gaul and, though
they were turned by Postumus as ‘Emperor of the Gauls’, the
old frontier territory east of the upper Rhine was never
again reoccupied by the Romans in force. Here are a few
of many occasions when in the 3rd century Free Germans
could have brought home great quantities of frontier loot
to their gods in their northern meres, just as in earlier times
the Romans had brought German booty home to their
Capitol. It may be added that there is some archaeological
evidence for the abrupt extension of the native culture of the
Danish islands to the Jutland peninsula about a.D. 200
(above, p. 29). But it is no more compulsory to postulate
that the bog-finds of Denmark represent great battles fought
in Slesvig or on Fyen than that the Roman triumphs of a
Germanicus or a Domitian imply victories in the Forum
Romanum. Whatever their precise and immediate origin,
the general significance of the peat-bog hoards is plain
enough as a picturesque footnote to 3rd- and 4th-century
history.
 V - FREE GERMANY: THE IMPORTS
N this section selected groups of Roman imports into
Free Germany are discussed summarily in respect of date
and distribution. No attempt is made to cover the whole
field; the maps and lists prepared by Dr. H. J. Eggers in
1951! have made any such attempt unnecessary for some
time to come. But samples of the coin-finds, metal-wares
and glass will serve to illustrate and amplify some of the
data collected in the previous sections, and will demonstrate
the richness of the field. We begin with a general indication
of the evidence of the coins.

a. COINS?

Between the Rhine frontier and Latvia, between the

Danube and Gotland, some thousands of Roman coins—
gold, silver and ‘copper’-—found their way into the soil of
Free Germany and Scandinavia from the time of Augustus
onwards. The circumstances of their finding and the
significance of their distribution have been discussed by a
number of scholars, notably by S. Bolin, whose conclusions
hold the field.* His detailed analyses are not here recounted,
but some of the more general aspects are necessary to our
picture, and are in fact implicit in certain of the previous
sections.
The coins occur singly and in hoards; of the latter more
than 400 are known, and not a few of the single coins may in
fact represent ill-recorded hoards. They are found in
graves, on settlement-sites, in peat-bogs, and at unrelated
find-spots. The extent nevertheless to which they circulated
as actual currency amongst the tribesmen was doubtless
1 See Bibliography, p. 182. 2 Pl. IX, p. 68, and fig. 6.
8 Fyndan av romerska mynt i det fria Germanien (Lund, 1926); résumé in
Bericht der Deutsches Arch. Inst. 1929 (Frankfort, 1930), 86ff.
64 ROME BEYOND THE IMPERIAL FRONTIERS

relatively slight. When we come to consider the Roman

coinage of India, the probability will emerge that there the
imported gold and silver coins were used largely in bulk, as
bullion guaranteed in quality and perhaps weight by the
imperial stamp, but rarely distributed as currency: a supposi-
tion reinforced by the absence there of Roman ‘copper’
coins before the 4th century, and by the deliberate mutilation
of Roman gold. The likelihood of a similar attitude in
inner Germany is supported by the statement of Tacitus
(Germ. 5) that ‘the Germans nearest to us value gold and
silver for their use in trade.... The peoples of the interior
employ barter.’ Consistently with this, in the time of
Tacitus, at the end of the Ist century A.D., copper coinage
was rare in Germany.1_ Later, on the other hand, presum-
ably as a result of commercial experience, this coinage
flowed more freely, and before the end of the Antonine
period (and century) was not infrequently included in graves
as far afield as East Prussia. After this time, copper must
have been used fairly widely as ‘small change’ in the normal
processes of monetary circulation, doubtless encouraged by
the fact that, with the deterioration of the silver after Trajan
and particularly after Septimius Severus, there was an
increasing approximation in metal-value between the two
categories.
That the quality of the Roman silver was carefully watched
by the German traders is clear from the further statement of
Tacitus that the Germans preferred ‘money that is old and
familiar, denarii with the notched edge and the type of the
two-horse chariot’ (pl. IX, p. 68). The barbarian habit of
adhering to certain well-known types of foreign coinage is
familiar enough, an example being the retention of Maria
Theresa dollars in modern Abyssinia. Even the Athenians
of the sth century B.c. found it politic to retain their archaic
coin-dies, and Celtic moneyers of the rst centuries B.c.-A.D.
adhered largely to a Macedonian type of the 4th century
1 Only 36 stray copper coins are recorded from West Germany and Holland
(outside the Roman frontier) of dates prior to the end of the rst century A.D.
 (IAXX ‘saipnig uvuioy fo jvunof) -oSe ‘ary 03 aotid AueursaD 99IJ Ut suTOD UO Y Jo uoNqINSI 9 ‘SI
I ae |
fF 6 SHBIW 320 BIVW9OS
66 ROME BEYOND THE IMPERIAL FRONTIERS

B.c. But the conservatism of Tacitus’s Germans was not

unreasoned; for in A.D. 63 Nero had reduced the silver-
content of the denarius and so enhanced the popularity of
previous issues in remote markets where intrinsic wort
rather than state guarantee was the basis of confidence.
Later, a similar ponding-back of pre-Severan denarii seems
to have occurred in the 3rd century after the further devalua-
tion of the currency by Septimius Severus. It may be
doubted, however, whether the imported coinage ever
entered significantly, even as bullion, into the domestic
trade of Free Germany; rather may we suspect that its
primary use to the German was as a medium for the purchase
of other imports from Roman or sub-Roman merchants,
of whose penetration into unconquered territories there are
many examples. In favour of this supposition is the scarcity
of pre-Neronic silver in German hoards later than A.D. 107
when Trajan finally called it in. Most of it must already—
or soon afterwards—have found its way back across the
Roman frontier.
In summary, then, it is to be supposed that in the peripheral
trade of Free Germany, where contact with traders from the
Roman world was recurrent, Roman coins circulated in so
far as their metal-value was intrinsically acceptable; in the
interior and the outlands they were used variously in bulk
as bullion, or individually as curiosities, gifts, ornaments,
and sometimes as funerary equipment; rarely as a local
instrument of exchange, since the Germans, unlike the
Celts, were unfamiliar with a monetary system and cannot
readily or widely have adopted this alien usage.
Roman gold coinage was unpopular in Germany prior to
the 4th century. Its value was presumably too high for
normal use. Tacitus (Germ. 5) observes that the Germans
of the interior tried to get silver in preference to gold; only
those ‘nearest us’ valued both gold and silver. The archaeo-
logical evidence confirms the statement. Reference has
been made (p. 54) to the scatter of Augustan gold within
1 Or, as pure silver, into the melting-pot.
 FREE GERMANY: THE IMPORTS 67
range of the Varus disaster of A.D. 9. Apart from this, a
small hoard of six aurei of Tiberius, Claudius and Nero has
been found at Tensfelderau, Kr. Segeberg, in Schleswig-
Holstein; and another of Tiberius is recorded from West
Germany. This modest total may be compared with up-
end
fom.
fo
"9
wards of 13 1st-century hoards of denarii in the same region.
©analy
remanent,

A hoard of aurei of the same date from Bohemia fits in with

the historical relationship between the Marcomanni and the
empire at this period. Other stray finds of aurei from West
Germany and Holland of dates prior to A.D. 100 all lie
sufficiently near the frontier to require no other explanation.
To the 2nd century a.p. no gold hoards are ascribed,
although occasional gold coins of that century, usually
pierced or otherwise adapted for ornament, penetrated to
Thuringia and Bavaria. Gold of the 3rd century in Germany
or Denmark was deposited mostly in the 4th century, when
Constantinian solidi took the field and were continued by an
increasing succession of gold deposits into the sth century,
particularly in Eastern Germany. There the fluid tribal
situation—particularly the Gothic and Hunnish migrations
—had set up a new current of intercourse between the Baltic
and the Black Sea, and new if superficial standards of luxury
were emerging under the influence of forceful contact with
the Byzantine world.
The distributions of Roman coinage in Free Germany
have already constituted a major basis for the section dealing
with routes and markets (pp. 11ff.) and, save by maps (figs.
t and 6), need not be amplified in detail for the general
picture which is the purpose of this book. It will suffice
to add that under the steadying influence of firm Imperial
rule hoards, the non-recovery of which is a sure symptom of
a time of trouble, are of rare occurrence in the earlier half
of the 2nd century; they do not again become emphatic
until the long period of disturbance which began with the
Marcomannic war of Marcus Aurelius in A.D. 167. That
war was itself merely a focus of unrest extending far beyond
the limits of Bohemia. For example, of hoards ending with
68 ROME BEYOND THE IMPERIAL FRONTIERS

coins of Marcus Aurelius, five are known from western

Germany, three from the mouth of the Vistula, others from
Posen, Schleswig-Holstein and Sweden. The Vistula hoards,
added to those of Commodus from the same region, may
represent a reaction to the earlier movements of the Goths
from their Swedish homeland on their long march towards
south-eastern Europe (p. 26). But the equation of particular
coin-groups with inchoate historical phases of this kind is
liable to be an arbitrary business and cannot be stressed.
So again a further suggestion, that a series of 3rd-century
silver hoards and gold coins, extending down to the emperor
Probus (A.D. 276-82), marks the migration of Burgundians
and Vandals from east-central Europe towards Gaul about
A.D. 278, is of questionable value. It is in fact possible to
isolate a number of these finds in an appropriate belt across
Germany through southern Brandenburg, Saxony and
Thuringia, but only by applying the blind eye to wide
variations in dating and distribution. Certainly without
the historical background no comparable movement could
fairly have been deduced from the coin-evidence as it stands.
On the whole, in so confused and turbulent a territory as
Free Germany it is best to restrict the use of coins primarily
to the analysis of main routes and concentrations, as an
amplification of other archaeological evidence, and above
all as a chronological control.

b. VESSELS OF SILVER AND GOLD

Free Germany has yielded an impressive quantity of
Roman. silver-ware, the combined product doubtless of
commerce, warfare and diplomacy. Pride of place must
be given to the great hoard discovered in 1868 at Hildesheim,
south of Hanover, where it had been carefully hidden in a
pit about 4 x 3 ft. dug to a depth of some 5 ft. below the
former level. Including fragments, it comprised about 70
pieces, and is one of the most remarkable collections of
silver known from the ancient world (pls. X-XII, pp.
 (Fo puv g ‘dd aag)
YILUUUOE WO SUTOD IDATIS ULIOYP API
IX
(See
69)
p.

dish
Silver
Hilde
the
from
hoard
Hanov
 FREE GERMANY: THE IMPORTS 69

69, 84-5). Most of the pieces must have been made

within the century 50 B.c.-A.D. 50 with a bias towards the
early Augustan period, but the name, Marcus Aurelius,
stippled on the base of one of them suggests, without
absolute proof, that they were not buried before the name
came into prominence in the 2nd century a.p. In the
circumstances, guesses as to the conditions of burial are
unprofitable. No relevant military or civil site has been
identified in the vicinity.
Amongst the contents of the hoard! were half a dozen
dishes, eight plates and trays, eight bowls, nine cups, two
urns, a bucket, four paterae, two tripod-tables and a candel-
abrum, all of silver, sometimes partially gilt: but whether
the total represented the luxury equipment of a general
officer’s ‘mess’ or whether it was otherwise assembled
cannot be deduced with any approach to certainty. It is
not merely in quantity that the hoard is outstanding, for two
at least of the pieces are unsurpassed amongst the surviving
examples of the Graeco-Roman silversmith’s craft. One of
these is a two-handled dish, elaborately ornamented with
honeysuckle and lotus patterns and with a seated figure of
Athena set as a central emblema in high relief (pl. X, p. 69).
Except for the arms, face and neck, the figure is (or was)
gilded, and certain elements in the floral design are also
picked out with gilding. The quality of the modelling,
particularly in the studied but not unrealistic grace of the
drapery, reflects the living tradition of Hellenistic sculpture
and gives the figure an independent value as a work of art.
Where it was made is disputable: Asia Minor and Alexandria
are two obvious competitors: but the artist was certainly a
Greek.
The other outstanding vessel is a dish with a central relief
or emblema representing the infant Hercules grasping two
snakes, framed by a tendril-pattern issuing from the tails of
two griffins and interspersed with birds (pl. XI, p. 84). The
pattern, the child’s clothing and the snakes were gilded.
1B, Pernice and F. Winter, Der Hildesheimer Silberfund (Berlin, 1901).
E
70 ROME BEYOND THE IMPERIAL FRONTIERS

There is evidence that the emblema is an insertion, and the

contrast of its skilled workmanship with the relative coarse-
ness of the tendril adds proof that the two parts did not
originally belong to the same vessel. The naturalistic child’s
head, rendered in the round, is a good example of late
Hellenistic genre, in contrast with the more academic
Athena of the dish described above.
Other pieces in this famous collection must here be passed
by with scarcely a mention: the dishes with emblemata
representing Attis and Cybele, bowls with the elaborate,
naturalistic floral patterns which delighted the Augustan
taste, urns with satyrs’ masks, herms and theatrical grotesques.
Three or four of the pieces show a coarser, less academic
handling which has suggested the possibility of a provincial
origin for them. Such are an oblong tray (pl. XII, pp. 84-5)
with vigorous representations of duck rising from the water
or diving below it, and, above all, two conical cups with bold
if somewhat uncouth friezes of animals, rosettes and leaf-
patterns. These may date from a time nearer that of the
burial of the hoard, and may be additions to it from a Gaulish
source. Unhappily, surviving examples of Gaulish crafts-
manship of the kind are too rare to offer satisfactory proof of
this attribution.
If we turn from the Hildesheim hoard to other finds of
silverware, we are at once on familiar ground; for of eleven
sites which have produced such ware, no fewer than nine are
those of ‘chieftains’ graves’ of the Liibsow group. The
scyphi of Hoby on the island of Laaland, of Dollerup in
Jutland, and of Liibsow I in Pomerania have already been
described (pp. 36ff.), and, with the exception of a minor cup
from Hoby, attributed to a Mediterranean (Graeco-Roman)
origin. Others may be ascribed rather to provincial work-
shops. A pair of plain but graceful cups from a grave at
Byrsted in Jutland has been regarded, without certainty or
perhaps probability, as Gaulish; but it is likely enough that
small silver bowls of late Roman date, vaguely imitating
cut-glass vessels, on the one hand from Haagerup in Fyen,
 FREE GERMANY: THE IMPORTS Up:
Nordrup in Zealand and Leuna in Saxony, and on the other
hand from Chaoucre, Dep. Aisne, were made in Gaul (pl.
VIls, p. 52. A provincial origin has been proposed for the
silver cups from Dollerup in Jutland (p. 40); and others from
Mollerup, also in Jutland, from the second grave at Liibsow
in Pomerania, from Leg Piekarski in Poland and from
Holubic near Prague’ all show un-Roman features, but have
not been satisfactorily identified with specific provincial
schools. Their normally excellent manufacture, the recurrent
use of gilding to emphasize the pattern, and their generally
classical types may be taken to indicate origins within the
Empire. There is at present no clear hint of any production
of this kind and quality within Free Germany itself. It
may be added that, with the exception of the bowls from
Haagerup, most of the silver vessels from Free Germany
belong by type and association to the first two centuries of
the Empire, with emphasis on the 1st century rather than
the 2nd. In other words, they largely precede the full
development of Romano-Gallic craftsmanship.
The region of the Black Sea remains at present an unknown
quantity in the luxury trade beyond the frontiers. It may be
conjectured that from or through this region came to Var-
pelev in Zealand an elaborate cup of blue glass with frame and
pierced foliage-overlay of silver, incorporating the Greek
greeting EYTYXW= (pl. VB, pp. 36-7). With the cup was a
coin of Probus (a.p. 276-82), and the burial may not in fact
be earlier than the 4th century, but we may recall that from
the latter part of the 3rd century south Russia was linked by a
more or less continuous Gothic zone with central Europe and
the Baltic (p. 26).2. A similar south-eastern origin has been
ascribed to the silver goblets from a well-known find at
Valloby in Zealand. The decoration of these goblets, in
particular a frieze of backward-looking animals below the
1 For these cups, see especially O. Voss in Acta Arch. XIX (1948), 252ff.
2 A 2nd-century date and an Italian origin have been ascribed to this cup,
and the occurrence of an heirloom in a late grave is not impossible. But I see
no reason why the cup should not be of the 3rd century, and the Eastern Medi-
terranean is a more probable source than Italy.
72 ROME BEYOND THE IMPERIAL FRONTIERS

rim, is consistent with a South Russian origin, and the

Goths may again be suspected as the carriers. Against this
supposition is the fact that amongst the associated objects
was a sigillata bowl signed by the potter Comitialis, who
worked at Rheinzabern in western Germany during the
latter half of the 2nd century, and this date, if accepted
literally, is earlier than that of the Gothic kingdom by the
Black Sea. Transit up the Danube valley and thence north-
wards along one or other of the well-trodden trade-routes
from Carnuntum is an alternative possibility which would
reconcile date and probable origin without involving the
Goths; and a similar path, at any rate from Carnuntum
onwards, may have been taken by an attractive silver cup, of
low cylindrical form with relief ornament representing
masks and a griffin bringing down a stag, found at Ostropa-
taka in Czechoslovakia. This cup is, however, of late date
and may equally well have passed through Gothic hands
beyond the frontier.
The only recorded gold vessel of classical type is a plain
chalice, about 6} in. high, found at Ostropataka with the
silver cup just mentioned, and presumably of similar origin
and date. The two famous golden drinking-horns from
Gallehus in Sleswig, though possibly of late Roman date,
have nothing to do with Roman craftsmanship or usage.

C. BRONZE WARE

Of bronze vessels, more than 850 have been found in Free

Germany (with Scandinavia), about half of them in Jutland
and the Baltic islands. The total is impressive, particularly
when it be recalled that sites other than graves have scarcely
contributed to it. Apart from the fact that relatively few
settlements of the period have been explored, the chance of
survival of a metal vessel on an occupation-site is very much
less than in a grave. At the same time, it is not to be
1 Conveniently accessible in H. Shetelig and H. Falk, Scandinavian Archaeo-
logy (Oxford, 1937), p. 208 and Pl. 36.
 FREE GERMANY: THE IMPORTS 73
assumed that the custom of burying these vessels with the
dead was coextensive with their total distribution. An
examination of the burials in the vicinity of the gulf of
Danzig has already demonstrated the contrary (p. 24).
Selected examples must here suffice to illustrate this great
mass of material. The series begins, so far as we are con-
cerned, with a number of imports dating from the end of the
Ist century B.c. and the first decade of the 1st century A.D.,
and may here be represented by two distinctive types: a situla
or wine-bucket with a pair of confronted dolphins at the
base of the handle-strap, and a saucepan or casserole with a
swan’s head on each side of the pierced end of the handle
(fig. 7). Of the former, six or seven examples of this
period are recorded: four or five from Bohemia and the
Thuringian border, one from the Oder valley in Branden-
burg, one from Mecklenburg near the Elbe, and one from
Hoby in Zealand—all with a single exception from crema-
tion-burials. Predecessors of the same general type had
found their way in the 1st century B.c. into the river-valleys
of the North German plain as far eastwards as the Vistula
and the Passarge, and one had even ventured as far north as
Sweden. Their ultimate origin was doubtless in Italy,
where Capua dominates the archaeological scene, but
whether at this early date they travelled northwards from the
Danube or found their way more circuitously through the
Celtic lands is not at present clear. The pattern becomes
unmistakable, however, in the first decade a.p. The
concentration of late Augustan types in Bohemia reflects the
régime of the romanizing Marcomannian king Maroboduus
and the recorded activity of Roman traders amongst his
tribesmen (p. 19). The old Aquileia~-Danube route was now
systematized via Carnuntum in the methodical Augustan
manner, and a reasonably regular traffic along it was the
natural sequel. Dolphin-buckets and swan-saucepans
travelled that way, and no fewer than twenty of the latter
have been recovered from Marcomannian graves, with an
outlier near the valley of the lower March as a pointer.
74 ROME BEYOND THE IMPERIAL FRONTIERS

Further north, the swan-saucepans reached the valleys of the

Oder and the Weser, but only one of them made Scandinavia
—a strangely isolated example from a grave at Kvale, far
up the Sogne fiord 80 miles north-east of Bergen. Only

ae
500
Heights over
Metres _——
SWANS-HEAD PATERA

| DOLPHIN SITULA IN” CENT. B.C. = x

Nae ; A.D. =0

My

Avalos Q__ 100 200 300 400 500 4; /ometres

Fig.7 Distribution of early types of Roman bronze vessels

(After H. J. Eggers)
 FREE GERMANY: THE IMPORTS 75

two of them approached the Baltic coast, and there is no

reason for associating the distribution of either type specific-
ally within the amber trade.
Other early Imperial types confirm these distributions:
for example, buckets or situlae with a human mask on the
attachments of the handle-loops such as that already described
from Liibsow (p. 44), paterae with ram’s-head handles, and
round-bottomed ladles or strainers with oar-shaped handles,
the bowls of the strainers pierced at first in the form of a
maeander. These imports extend the map-pattern slightly
towards Poland and Sweden, but the main bulk of them
remains concentrated in central Europe and Denmark. Italy
—in particular, Campania—was still the source of supply,
and the Carnuntum route remained the principal approach.
As the Ist century wore on, however, a variation began to
emerge: after the flight of Maroboduus in a.p. 18 (p. 21),
Bohemia, home of the Marcomanni, gradually lost the
commercial impetus which the king had given to it, and was
increasingly by-passed by the Italian trade with the north.
An illustration will make this clear. Partially contemporary
with the swan-handled saucepans already mentioned, but
outlasting them, were other types marked by a swan-less half-
round or circular piercing through the handle. These
scarcely occur in Bohemia, but are abundant further north,
particularly in the Danish islands. Eggers notes that, of
those with the circular piercing, only two have been recorded
from Bohemia as against no fewer than 51 from Denmark,
mostly from inhumation-graves (fig. 8). The map-pattern,
now focused on c. A.D. 50 and later, still suggests a Danubian
rather than a Gaulish or North Sea approach, but against this
must be set the cessation of the productive graves towards the
west. In the circumstances, it is, as often, impossible, in the
interpretation of the map, to estimate the extent to which the
distribution of a burial custom may distort the archaeological
distribution. One other early type calls for remark even in a
summary list. This is a plain convex-sided bronze pail
about 9 in. high, with simple eyelet for the handle, known
76 ROME BEYOND THE IMPERIAL FRONTIERS

as the ‘Eastland’ type. Ten of these occur in the territory

of the Marcomanni, whom they presumably reached in the
first quarter of the rst century a.D. For the rest, they cover
what we may now term the conventional early 1st-century

peel Heights over

500 Metres
co}

/00 300 400 500 Kilometres

Fig. 8 Distribution of Roman paterae with circular-piercing

(After H. J. Eggers)
 FREE GERMANY: THE IMPORTS TH

zone in central Europe and Denmark; but the type lasted on

well into the 2nd century and perhaps even later (p. 42),
and, unlike the contemporary types already discussed, spread
across into Norway and Sweden, where, excluding the

Heights over
© 500 Metres

9190
290 300 400500 Kilometres

Fig. 9 Distribution of ‘the poor-man’s bucket.’ (After H. J. Eggers)

 78 ROME BEYOND THE IMPERIAL FRONTIERS

Swedish islands, over 50 examples have been found (fig. 9.)

In Norway they are probably all derived from cremation-
graves, in which they conveniently contained the ashes. In
Sweden they occasionally occur also as grave-furniture
with inhumations. Why this unusual spread into the far
north?
The reason which has been given by more than one writer
is doubtless the correct one: these vessels are the simplest and
cheapest of their kind, and were alone within the purchasing
power of the poorer folk of the hills and fiords beyond the
Baltic; only the relatively wealthy aristocracy of Denmark
and the German Plain could afford elaborate and costly
bronzes such as those of many of the chieftains’ graves.
The differential factor here, therefore, was an economic or
sociological one, and had little or nothing to do with burial
custom or trade facilities. The archaeological lesson is
worth noting.
By the Later Empire—the 3rd and 4th centuries—changes
had taken place in the orientation of the traffic in bronze
vessels, as of other commodities, though in bulk the trade
showed little change. Save for the occurrence of two types
of bucket or situla (Eggers, maps 17 and 20), Bohemia was
- commercially no longer on the map. Again with these
questionable exceptions, trade with Italy had been replaced
by a considerable trade with Gaul and an intermittent trade
with the Black Sea region. Denmark, and in particular
Zealand, still remained a major customer, but now Thuringia
and the lowlands between the Elbe and the lower Rhine
were substantial rivals, whilst Poland is occasionally repre-
sented and would doubtless make a more impressive showing
if more adequately explored.
To illustrate these new trends, no better example can be
taken than the so-called “Hemmoor’ buckets (fig. 10).
These are simple handled bowls 7-9 ins. high on a footstand,
with a linear or, occasionally, figured zone below the rim.
They occur along and behind the Rhine-Neckar-Danube
frontier, not infrequently in hoards, which have been
 FREE GERMANY: THE IMPORTS 79

associated with the break-through of the Alamanni in these

parts about a.p. 258. Their Gaulish or West German origin

Heights over
500 Metres

Fig. 10 Distribution of ‘Hemmoor” buckets. (After H. J. Eggers)

is certain, and their manufacture has been variously ascribed

to Cologne and to the neighbourhood of Aachen; an
80 ROME BEYOND THE IMPERIAL FRONTIERS

inscribed example from as far afield as Sweden may have

come from Raetia (p. 53). An analysis of their distribution
beyond the frontier suggests interesting interpretations. In
Denmark and Mecklenburg they are found largely with
inhumation-burials, and the same rule applies to Thuringia;
in Lausitz they are associated with, but do not contain,
cremations, and are often ‘ritually broken’ in accordance
with a widespread funeral custom. From the mouth of the
Elbe to the lower Rhine they occur also in cremation-graves,
but here as containers of the burnt bones. How far these
variations may be equated with tribal groups is a dangerous
speculation which must await some future analysis of the
associated native cultures. In any event, the task is com-
plicated almost beyond hope by the fluid condition of the
German tribes in the 3rd century; and to suggest, as has been
done, that the Lausitz folk were Burgundians and the
western group Saxons is to outpace both the historical and
the archaeological evidence.
These vessels, and others of the same phase—straight-
sided bronze basins, late bronze strainers with oar-shaped
handle and flat base (Eggers types 78-88 and 161), were
doubtless distributed by the standard routes: the tributary
valleys of the Rhine as far south as the Main, and the North
Sea coast extended by a land-traverse across the base of the
Jutland peninsula. It is worthy of note that five examples
reached the southern coast of Norway.
Mention of Norway may remind us that, except for the
islands, Norway and Sweden have mostly been peripheral to
our study. Now, however, in the 4th century, on one
occasion they dominate the scene. A group of bucket-
handled bronze vessels with cylindrical neck and sharply
rounded or carinated body (Eggers types 11-14, “Westland
type’) occurs in 19 cremation-burials in Norway and
Sweden, reaching as far north as the Trondheim fiord on the
west and Sundsvall on the east; the latter point may have
been reached by a land-route across country from the former
(fig. 11). This remarkable distribution is emphasized by the
 FREE GERMANY: THE IMPORTS 81

fact that only four examples are known from Free Germany
south of the Baltic—one each from inhumation-burials on
Fyen and in Mecklenburg, one from a cremation at the
mouth of the Elbe, and one from a peat deposit by the middle

pees Heights over

2 500 Metres we
oG”

AS

Scale: (a) 100 200 300 400 500 Ki'Jometres

Fig. 11 Distribution of bronze vessels of “Westland” type

(After H. J. Eggers)
82 ROME BEYOND THE IMPERIAL FRONTIERS

reaches of the river. Owing to the general absence of

associated goods from the northern graves, the chronology
of these vessels is largely theoretical, though the late Roman
dating is fairly well established by the two inhumation-
burials. They are later, perhaps considerably later, than
the ‘Hemmoor’ type, and may well, as Eggers suggests,
spill over into the sth century. A similar suspicion arises
in regard to other late groups of bronze or glass with
a comparable though less isolated distribution (p. 30).
If so, they are a not undramatic expression of coastal
concentration and activity on the eve of the great Folk-
wandering. But the evidence is at present of a sketchy
character.
Traffic with southern or south-eastern Europe, in particular
the Black Sea region, has already been illustrated tentatively
(p. 24) by reference to certain types of fluted bronze vessel
which, rather by process of elimination than by direct
evidence, have been ascribed alternatively to that source or to
Italy (fig. 12). Of the two main types of fluted bowl, the
earlier, which seems to emerge a little before a.p. 200, has a
rounded profile, fine fluting and a twisted handle with attach-
ment-loops flanked by swans’ heads. The type-site is Gile,
near Oslo, and the type is, curiously enough, relatively
common in Norway, though it also occurs on eight sites in
East Prussia and Pomerania. The later variety has upright
sides, coarser fluting and a plain handle, and is known as the
Valloby type from a site in Zealand. This also appears in
the lands beside the lower Vistula. Along the Gaulish
frontier the general type occurs rarely and only in circum-
stances which can be explained by garrison-movement from
central or eastern Europe. In Italy the type is not known,
but our knowledge there of post-Pompeiian metalwork is
slight. It may occur on the margins of the Black Sea, but
these are at present barred to research. The most certain and
significant feature attending it is its presence in some force in
East Prussia, northern Poland and eastern Pomerania, a
region approached from the south-east by the river-arteries
 FREE GERMANY: THE IMPORTS 83

of the Vistula and the Dniester; and Italy is perhaps less likely
as the starting-point than are the obscured markets and

Heights over
500 Metres

Scale: Q 190 200 =300 400 —~500 Ki'/ometres

Fig. 12 Distribution of fluted bronze bowls. (After H. J. Eggers)

factories of south Russia, linked thus readily with the north.
Such is the view of the latest investigator (Eggers), though his
84 ROME BEYOND THE IMPERIAL FRONTIERS

principal predecessor (Ekholm) preferred Italy.1 There the

matter may be left until the door is open for further research.

d. GLASS
Under the Early Empire, glassware was not traded
abundantly into Free Germany; some 14 pieces in all can be
ascribed to the 1st or early 2nd century a.p. Of these the
most abundant type is that of a ribbed or pillared bowl], such
as may be found throughout the vast zone covered by Roman
trade at that time, from Britain to India (fig. 19, 1-2). Eight
illared bowls have been found, ranging from Oslo to
Poland, all probably or certainly in ‘chieftains’ graves’ of
the Liibsow group (p. 32). (It may be recalled that some
of the ‘Chieftains’ were women: a particularly fine bowl, for
example, of white-marbled blue glass was recovered from a
woman's grave at Juellinge on Laaland—see pl. IV, 3, p. 36
—together with other imports, including a bronze patera
signed by the well-known maker L. Ansius Diodorus, who
worked at or near Capua between c. A.D. 60 and 90.) The
remaining half-dozen early glasses are conical beakers, either
cut to facets or painted. An example from one of the
Liibsow graves bears two painted zones of eladiators.
All these glass vessels probably came from Campania by
way of Carnuntum and the central European river-system.
The factories of provincial Germany belong to the Middle
and Later Empire, the glass of which has recently been
reviewed in detail by Eggers, with inevitable reservations
as to the part played by South Russia in 3rd- and 4th-
century distributions. His maps show a great concentration
in the Danish islands, especially Zealand, a fair scatter along
and behind the Baltic coast as far as the eastern end of the
Gulf of Danzig, a minor concentration in Thuringia,
sporadic occurrences far to the south-east along the Dniester
and in the Ukraine, and a notable extension towards the
north in Sweden and Norway, as far north indeed as the
1 Acta Arch. VI (1935), 7Uff.
 XI

Silver dish from the Hildesheim hoard, Hanover. tof

(See p. 69)
Silver tray from the Hildesheim hoard, Hanover. 2
(See p. 70)
XIII

sslan
eghs

south
Ru

ssibly
of
86)
(See
p.

Zealand.
Nordrup,
from
,
beaker,
lass
po
c origin.
2S

B.

)
1

Denmark.
|erg,

Thorsb

58)
(See
p.

the
moss
rom

frtrousers
en

Wooll

INS
 XIV

A. Painted glass cup of ‘Nordrup-Jesendort’ type, from

Himlingoje, Denmark. ? (See p. 85)

B. Cup of green glass with ‘claw’ decoration, from Vedsted, Zealand. 2 (See p. 86)
 FREE GERMANY: THE IMPORTS 85
Trondheim fiord. On the other hand, between the Elbe
and the Rhine, except in Thuringia, only two determinate
glass vessels of the late period have come to light, in spite of
the relative proximity of the Rhenish factories. The
explanation offered is doubtless in part a right one: that in
the west the dominance of cremation has robbed the archaeo-
logist of any easy source of material. Against this, however,
must be set the fact that cremation was likewise prevalent in
Norway, which has nevertheless proved singularly produc-
tive. Chance and archaeological skill or enterprise may be
suspected as contributory factors. Paradoxically, Roman
glass is completely absent from Bohemia, destined long
afterwards to become itself a European centre of glass-
manufacture. During the brief early period of commercial
prosperity there, glass was rarely traded across the frontiers;
and later, when glass became a major Roman export,
Marcomannic trade had dwindled almost to extinction.
Notable amongst late types which can be ascribed to
western factories are the somewhat rare cylindrical cups,
painted with human figures and those of animals, to which
the name ‘Nordrup-Jesendorf’ has been given from find-spots
respectively in Zealand and Mecklenburg (pl. XIVa, p. 85).
Nineteen examples of this type have been found, representing
eight localities, all of them accessible by the North Sea
coastal route from the Rhine, with a land-traverse across
the Danish peninsula. Another type which can be recog-
nized as western by form and fabric is a small bowl with
applied ribs; a type represented mostly in Thuringia with an
outlier east of the Oder. Other small bowls with cut
geometrical patterns, a type characteristic of Cologne, occur
not only in the Danish islands, Mecklenburg and Thuringia,
but also far away amongst the tributaries of the Danube and
the Theiss and northwards by the gulf of Danzig, a distribu-
tion which suggests in part a sea-route and in part a cross-
country route from the upper Rhine to the Danube. We
have seen that these cut-glass bowls or cups were sometimes
imitated in silver (pl. VIIB, p. 52). Other types from Free
G
86 ROME BEYOND THE IMPERIAL FRONTIERS

Germany are noted by Eggers as foreign to the Rhenish

factories and, with some help from geographical distribution,
are ascribed provisionally to the Black Sea markets, though
more positive evidence is admittedly required. Relatively
tall, slim beakers with applied thread-ornaments, differing
from their western counterparts both in form and in the
colour and quality of their threads, occur in Zealand (pl.
XIllz, pp. 84-5, in the neighbourhood of the Vistula delta
and in the valley of the Dniester, and are thus attributed.
They seem to belong mostly to the 3rd century. Two other
south-eastern or at least non-western types have a further
interest in that they bring Norway and Sweden into the
picture, presumably by way of the Dniester, San, Vistula
and Denmark. The first consists of small glass cups with
claw-like ribs extending from the base halfway up the
sides (pl. XIVB, p. 85). This late type has been found on
three sites near the mouth of the Vistula, on Bornholm and
Zealand, at the north end of Jutland, on two sites in southern
Norway and one site in western Sweden. The other type
comprises tumblers of thick olive-green glass with cut orna-
ment in which sunk ovals predominate (pl. XVa, p. 100).
In date these tumblers come at the end of our period and
overlap that of the Folk-wandering, which their distribution
in Norway (17 sites) seems to herald or reflect. In this
respect they conform closely in time and place with the
‘Westland’ type of bronze bucket discussed above (p. 80).
For the rest the tumblers occur on two Swedish sites, in
Jutland and the Danish islands, in Pomerania, at the mouth
of the Vistula and in the upper valley of the Oder, thus
conforming with the familiar pattern of south-central and
south-eastern Europe. Incidentally, a factor in the wide
distribution of these tumblers may have been the unusually
thick and durable glass of which they are made.
Lastly, a word must be said of the glass drinking-
horns with which Rhenish factories supplemented the
metal-mounted ox-horns of the German tradition (p. 21).
Whether the Rhenish factories exercised a monopoly in
 FREE GERMANY: THE IMPORTS 87

this matter is less certain. Glass horns have been found

in inhumation-graves in Southern Norway (two), Sweden
(one), Jutland (one), Zealand (two), and East Prussia (one),
and appear to date mainly from the 4th and sth centuries,
perhaps with emphasis on the later date (pl. XVz, p. 100).
The period is one in which, on present knowledge, it is
increasingly difficult to allot sources as between east and
west, and the distribution is too generalized to help. Once
again, the Iron Curtain bars the necessary research.

€. TERRA SIGILLATA

The ease with which the familiar red-glazed pottery known

as terra sigillata or ‘Samian’ can be recognized and roughly
dated gives it a special value to the archaeologist, and the fact
that 262 sites in Free Germany and Scandinavia have
produced sherds of it! brings it prominently into the present
context (fig. 13). Let it be said at once, however, that a large
proportion of these sites represents short-range trade to the
Frisian coast or up the tributaries of the Rhine. Friesland
alone accounts for not less than 42 sites; and the limitations
of the trade are emphasized by the circumstance that
Denmark, in spite of a century of intensive exploration, has
yielded only four examples, whilst on the Swedish island of
Gotland, rich in Roman bronze vessels and not lacking in
Roman glass, only one bowl of terra sigillata has been
recorded. This solitary Gotland bowl, incidentally, offers
a warning to the student. It comes from an ancient farm-
settlement at Kinne, on the south-eastern shore of the island,
where systematic excavations were carried out in 1926-31
and produced 39 sherds of decorated sigillata spread over an
area of about 400 sq. m. The fragments cannot be made to
join one another, but have been shown to represent a single
bowl. It is a cautionary thought that a less careful in-
vestigation or a less distinctive vessel might easily have
1 Dagmar Selling, ‘Terra Sigillatafynd i det Fria Germanien’, in Kultur-
historiska Studier tillagnade Nils Aberg (Stockholm, 1938), pp. 101-114; and a
fuller list in Eggers (1951), p. 182.
88 ROME BEYOND THE IMPERIAL FRONTIERS

multiplied the distribution-density of terra sigillata on

Gotland 39 times!

“s Heights over
“= 500 Metres

2190
200 300 400 500 Kilometres
Fig. 13 Distribution of terra sigillata. (After D. Selling and H. J. Eggers)

The earliest sigillata found in Free Germany is an Arretine

vessel (probably early 1st century a.p.) from Holtgaste, Kr.
 FREE GERMANY: THE IMPORTS 89
Leer, in Hanover. The extreme scarcity of Arretine or
Italian ware in Free Germany is, however, in contrast with
its comparative abundance in Britain (at Colchester and
elsewhere) before the Roman conquest. A few specimens
of early provincial sigillata, made in the south of Gaul
between a.D. 25 and 95, have been recorded: notably a
bowl of the type known as ‘Dragendorff 29’ from Vippa-
chedalhausen, Kr. Weimar, in Thuringia, and a dish of form
‘Dragendorff 18’ stamped oF Bassi Co (i.e. the factory of
Bassus and Coelus) from Mehrum near Diisseldorf, and
therefore not far from the Roman frontier. Both of these
vessels date approximately from the third quarter of the 1st
century A.D. Another bowl, made probably at Banassac in
South Gaul towards a.p. 100, has been found at Dzwinogrod,
Krets Bobrka, in Poland, and is the most easterly example for
which a Ist-century date can be claimed. It may have
found its way up from Carnuntum on the Danube rather
than transversely across north-central Europe.
For the rest, the sigillata now in question dates mostly
from the 2nd century and was derived either from Lezoux in
central Gaul or from one or other of the factories which
sprang up in the frontier territories from the end of the ist
century A.D. onwards: at Heiligenberg, Rheinzabern,
Blickweiler, Trier and Westerndorf. Thus, one of the four
Danish examples (from Valloby in Zealand) is a bowl
stamped by Comitialis, who worked at Rheinzabern in the
middle and latter part of the 2nd century; a decorated bowl
by Cintugnatus, who was at Heiligenberg in the first half
of the century, is recorded from Borstel in Saxony; another
bowl (form 37) from Dortmund in Westphalia bears the
name of the Trier potter Amator, of the second half of the
2nd century; Tarvagus of Westerndorf, mid-2nd century,
stamped a vessel from Dambitzen, Kr. Elbing, in West
Prussia; sherds from Ciosmy (near Lodz) and Dzwinogrod in
Poland bear the names of Elenius and Cintusmus, who
worked at Rheinzabern and Westerndorf after the middle of
the 2nd century; whilst as long ago as 1776 a bowl of form
90 ROME BEYOND THE IMPERIAL FRONTIERS

37 was recorded from Gispersleben, near Erfurt in Thuringia,

bearing the name of Drusus, who worked at Lezoux in the
first half of the 2nd century. A few examples of sigillata
are decorated in barbotine and ‘cut-glass’ techniques, which
may carry them into the 3rd century; but, save for inferior
derivatives, the manufacture of this fabric ceased after the
first quarter of that century.
If we turn from date and type to distribution, certain
points of interest arise. First, the terminal character of the
trade with the Frisian coast and its hinterland, extending to
the mouth of the Elbe, is sufficiently clear from the abundance
of sigillata on the terps and its absence from the Danish
peninsula. It is evident that the Frisians were in this respect
absorbers rather than mere carriers. Secondly, the trans-
fusion of wares into the valleys on the eastern flank of the
middle Rhine may mostly be dismissed as petty frontier
trade, but, on the other hand, some part of this traffic went
further afield: up the Lippe, for example, to the Minden-
Hameln crossings of the Weser, and along the Main-Fulda
and Wetterau routes to the vicinity of the Thuringian
Forest, where nearly twenty sigillata find-spots recall the
considerable concentration of Roman bronze vessels and
coinage hereabouts (p. 22). Thuringia indeed appears once
more as a rather mysterious island of romanization in
partibus. Thirdly, reference has been made to some slight
evidence—too slight at present for emphasis—that sigillata
occasionally reached east-central Europe and the Baltic via
the amber-route from Carnuntum. Some of the German
provincial potteries, notably Westerndorf, traded with the
Danube frontier, and sigillata sherds found in Czechoslovakia
(at Prodbaba, near Prague), in East Prussia near Elbing at the
mouth of the Vistula, and, less certainly, the Gotland (Kanne)
bowl already mentioned have been attributed to Westerndorf
and are perhaps more likely to have travelled by way of the
Danube and thence northwards than more directly eastwards
from the Rhine.
| VI : FREE GERMANY: SUMMARY

‘|= previous sections have shown, if selectively, the

remarkable extent to which Roman things penetrated
into Free Germany between the 1st and 4th centuries
A.D. From that summary much has been omitted; brooches,}
for example, and pins and beads, in all of which native
elements (Gaulish, Danubian, German) are mingled almost
inextricably with other elements from the south and west.
But detailed studies by categories or localities are not the
function of this book and have indeed been supplied in liberal
measure for the time being by the lists and maps of Drs.
Bolin and Eggers.? Here it must suffice to observe the general
trends of the evidence which they and others have collected.
That evidence points to three main and in some degree
successive sources. The first source was Italy, which in the
Ist century A.D. and well into the and century was the manu-
facturer and primary distributor of most of the Roman
goods found on Free German soil. Wares travelled by an
adequate road-service to the Danube and thence by the
amber-routes to the Baltic or, in the early part of the 1st
century, to Bohemia. Here the Marcomanni were for a
time active both as ultimate customers and as middlemen
in the central and northern European trade, but after the exile
of Maroboduus fell quickly out of the running. In the
west, the provinces of Gaul and western Germany absorbed
Italian goods and passed them on to barbarian buyers by sea
and land. Trade was supplemented by diplomatic gifts to
native princes, who, incidentally, sometimes reciprocated;
as when the Cimbri sought to propitiate Augustus by the
gift of a peculiarly sacred vessel (Strabo, 7. 2. 1). War
booty must also have contributed to the interchange, but is
difficult to isolate. What the Roman traders received in
1 For these the standard work is still O. Almgren, ‘Studien iiber nordeurop-
eische Fibeleformen’, Mannus, nr. 32 (Leipsig, 1923).
2 See Bibliography, p. 182, and p. 63, footnote 3.
92 ROME BEYOND THE IMPERIAL FRONTIERS

return for their enterprise is almost unknown. Slaves may

be assumed to have formed a part of the apparatus, although
the human fruits of victory must have constituted a serious
rival source. Skins may also have been included; Drusus
exacted from the Frisians a tribute of ox-hides for the
equipment of his army (Tacitus, Ann. IV, 72), and these same
Frisian ranchers doubtless continued to supply hides or
livestock to the neighbouring provinces. Marcus Aurelius
received cattle and horses from the Quadi and others (Dio,
LXXI, II). From Scandinavia, furs and dried fish were
probably traded; furs at least were widely used for clothing
by the Germans, although we have no contemporary
evidence for them in the present context. Amber from
Jutland and the coasts of East Prussia remained throughout a
primary commercial objective. But, with all allowance for
these obvious commodities, there is much in the apparatus of
Romano-German commerce that is unknown to us and is
likely to remain so. Perishable imports to Italy and the
provinces must have formed a major part of it.
Attention has been drawn to the early prominence of the
islands and mainland of Denmark in this traffic, and it has
been suspected that their prominence reflects some measure of
political cohesion, probably with Zealand as its focus.
Whether, or how far, piracy may have contributed to this
accession of wealth in the western Baltic can only be guessed,
but in any case some more regular form of exploitation—
doubtless including control of amber agencies—must be added.
How far Italy remained a source of supply after the earlier
part of the 2nd century is difficult to say. Our knowledge
of Italian factories in the Middle and Later Empire lacks
precision, and the routes across the Danube may have con-
tinued to contribute more than we know. Certainly,
however, by the end of the 2nd century the second main
source of supply had come into operation: Gaul and the
Rhineland. Pottery, glass, metalware and coinage now
began to flow thence eastwards and north-eastwards, en-
couraged alike by prosperity in the west and by central
 FREE GERMANY: SUMMARY 93

European wars and migrations which, from the time of

Marcus Aurelius, must have slowed down such direct
Italian traffic as remained. Here and there an attempt has
been made to trace the impact of tribal movement in the
map-pattern of Roman commodities, especially coin-hoards,
but with very uncertain success. The astonishing bog-
hoards of the 3rd and 4th centuries are presumably, how-
ever, the product of victories or successful forays on the
fringes of the Empire, and again emphasize the dominance of
the Danish lands.
The third main source is assumed without close docu-
mentation to have been based on Graeco-Scythian South
Russia, with extension to Byzantium. Much further
research, not at present feasible, is required before the share
of the factories and markets of this region can be adequately
appraised. Meanwhile, it is rather on negative than on
positive evidence that certain categories of glass and metal-
work from Free Germany are ascribed to the link established
by the Goths in and after the 3rd century between the Baltic
and the Black Sea. This link was later supplemented by the
impact of the Goths and Huns upon the Empire itself. The
phase lies outside the scope of the present book; otherwise it
might have been fitting to distinguish Byzantium as a fourth
source rather than as a supplement to the third. The great
quantities of Imperial gold which now found their way to
the southern shore and eastern islands of the Baltic indicate
the trends of the time. It may indeed be that, when the
contemporary historian Priscus affirms that Attila the Hun
ruled ‘the islands in the Ocean’, those of Bornholm, Oland
and Gotland are intended,1 and that the numerous gold
solidi found there represent the ultimate diffusion of plunder
or tribute from the Byzantine Emperor. For example, in
A.D. 435 Attila was able to impose upon the Eastern Empire an
annual tribute of 700 Ib. of gold, and seven years later the
annual levy was increased to 2,100 lb. Established by such
means on a vicarious Gold Standard, east-central and Baltic
1 See E. A. Thompson, A History ofAttila and the Huns (Oxford, 1948), p. 76.
94 ROME BEYOND THE IMPERIAL FRONTIERS

Europe poised glittering upon the brink of one of the

dramatically formative periods of history.
A final comment may be added. What on the broadest
and longest view, was the historical significance of this
penetration of wares, often of a high grade, far into the
interior of barbarian Europe during more than four centuries?
Is the study of them merely the laborious collection of
disjecta, of cultural accidents and curiosities out of context?
I think not. They played in fact a very positive part in
history. They are milestones on the barbarian road to El
Dorado. The subsequent literature of Northern Europe
makes abundantly clear the fascination which the Empire
exercised upon the minds of those without it. The Empire
was the Promised Land, of wealth untold and of perennial
allure. Large-scale migration is commonly attributable to
economic and social factors based on climatic change, soil-
exhaustion, unbalanced population, changing markets,
political encroachment. But to these propelling causes
which lie behind migration must be added others which lie
in front of it, attract it or stimulate it. To barbarian Europe,
at least from the middle of the and century onwards, the
Empire was in the latter category. Marcomanni, Franks,
Alamanni, Burgundians, Vandals, Goths, whatever in-
voluntary rearward forces propelled them, at least knew well
where they were going. Traders and raiders and perhaps
returning conscripts or mercenaries had already demonstrated
the way to fortune. The commodities which we have
catalogued marked out the highroads of great folk-move-
ments and advertised along them the delights which lay
ahead. Only when at the end of the 4th century the Huns of
Asia entered abruptly into the European scene were the main
lines of barbarian enterprise bent rudely from the traditional
objective. A wedge was now driven deeply into European
migration; and thenceforth the tribes of northern Germany
and Scandinavia, cut off from the sun, turned their faces
with a new determination to the mists of the North Sea.
We leave them there and turn our own faces to the far south.
 PAK -APRIGA

eer

AD ated Be date We BaNeat

N recent years the Roman occupation of Algeria and
Libya has been the subject of illuminating researches
carried out on the ground and from the air by French,
Italian and British explorers These researches require
co-ordination and extension, but meanwhile our picture of
the organization of the provinces of Numidia, Tripolitania
and Cyrenaica has acquired a new actuality, and coherent
maps ate for the first time available. The present is not the
context for a review of this evidence. It will suffice here
to indicate certain general features by way of introduction
to a more ample notice of discoveries in the deserts beyond
the frontier zone.
The mountainous geography of Algeria diversified the
settlement of the countryside to a greater extent than did the
broader natural partitions of Libya, and at the same time, by
canalizing traffic, invited a more definitive system of frontier
control. That control was based, perhaps from the time of
Hadrian in the twenties and thirties of the 2nd century a.D.,
on an intermittent artificial barrier of varying design some
500 miles long, linking natural obstacles of mountain and
marsh and cutting the approaches into the province from
the south-west. A buffer zone was added or organized by
the Emperor Septimius Severus, himself of African origin,
through the establishment of coloni (controlled cultivators)
outside a part of the frontier line about a.p. 200; and apart
from this specific act there is no doubt that a large share in the
1 See particularlyJ.Baradez, Fossatum Africae (Paris, 1949); R. G. Goodchild
and J. B. Ward Perkins, “The Limes Tripolitanus in the Light of Recent Dis-
coveries’, Journ. Roman Studies, XXXIX (1949), 81ff.; R. G. Goodchild, ‘The
Limes Tripolitanus : IP, ib. XL (1950), 30ff.
96 ROME BEYOND THE IMPERIAL FRONTIERS

defence of the frontier was still delegated as late as the

beginning of the sth century to the native farmers, who, as
remains a fields and irrigation works show, were settled
abundantly in its vicinity. ” Details of this native occupation,
whether inside or close outside the barrier, are at present
inadequately known and in any event do not concern us here.
Further east, however, the situation was at the same time
clearer and less closely defined. Libya falls naturally from
north to south into three or four roughly parallel zones:
the coastal plain, the broken rocky plateau or gebel, and
a belt of partial desert best designated as steppe, merging
finally into the variable landscape of the Sahara. The first
was inevitably the land of the considerable colonial cities,
Carthaginian or Greek in origin, later enlarged and supple-
mented by Imperial commerce. The second vee third were
the lands of semi-nomadic pastoralists and cultivators, whom
it was a task of Rome to anchor both as producers and as
a first-line militia. The fourth was the limitless reservoir of
nomadic enemies to all that the settled societies of the Medi-
terranean stood for. Had a fixed frontier barrier, a wall ora
ditch, been envisaged, it must have stood between the third
and fourth of these zones. But the absence of easy tactical
command on the fringe of the desert would have made a huge
line of this sort both costly to build and impossibly costly to
maintain and defend. Instead, the frontier zone was patrolled
by occasional vexillations or detachments of the Third
Augustan Legion (until its disbandment in a.p. 238) and, at
least from the beginning of the 3rd century, by bands of local
irregulars. Under the Later Empire, however, defence
crystallized primarily into a static complex of fortified farms
or gasrs in which the farmers of the gebel and the steppe
increasingly barricaded themselves and so, incidentally and
somewhat uncertainly, barricaded the province. Analogy is
presented partially by the Scottish brochs or fortified farmsteads
of the Romano-British period; more closely, by the pele
towers maintained by the yeomen of the Anglo-Scottish
border-lands in the Middle Ages.
 THE SAHARA 97
Through this miscellaneous frontier belt the cities of the
coastal fringe received, in varying degrees, the products of the
African interior. Ivory, precious stones, gold-dust, ostrich
feathers, slaves are mentioned or inferred, and, above all,
animals for the amphitheatres of Rome and elsewhere.
Some of these commodities were obtainable within the
provinces themselves, but some must have come from further
afield, transported by the caravans which were the principal
asset of the desert nomads. A French historian has indeed
remarked that, if Amsterdam be said to have been founded on
herring-bones, Lepcis Magna, the great trading city on the
Tripolitanian coast, may be said to have been built on the
carcases of camels. The proviso should be added that there
is no certain evidence for the normal use of the camel in the
coastal tracts before the 3rd century a.p.

i. THE FEZZAN

The Roman markets most accessible to this long-range

caravan trade were the three which gave Tripolitania (Land
of the Three Cities) its name: from west to east, Sabratha,
Oea (now the town of Tripoli) and Lepcis Magna itself, all
set in a southward re-entrant of the Mediterranean. It was
no accident that the hardy column which, under General
Leclerc, struggled through from Lake Chad to the sea in
1943 made the coast at Tripoli. Whether the west coast or
central Africa were the ultimate goal, the traveller from the
Mediterranean would usually make his way southwards to
the east-west ridge which forms the spine of the Sahara from
the Hoggar oa the west to Tibesti on the east, and would
tend to travel along it, one way or another, rather than
engulf himself in the boundless sand-plateau at its foot.
And, appropriately enough, midway between that ridge and
the coastal plain, in the Fezzan 430 miles (in a straight line)
south of Lepcis Magna, a remarkable assemblage of Roman
things was brought to light in 1933-4 by an Italian mission
led by Professor Pace in association with Dr. Caputo and
98 ROME BEYOND THE IMPERIAL FRONTIERS

Professor Sergi! The mission was in the field for little

more than three months, and its results testify alike to
the energy of its directors and the abundance of the
material.
The region in question lies some 50 miles north of
Murzuch, one of the capitals of the Fezzan, and is itself
bounded on the north by the Wadi el-Agial, which runs
east and west by way of Germa, the ancient Garama, known
to Pliny as ‘the celebrated capital of the Garamantes’ (V, 36).
To the north of the wadi (seasonal watercourse, dry for the
greater part of the year) lies the sand-sea, practically devoid
of life; to the south a strip of stony plateau, the hammada,
rises to a range of hills, and offers a milieu suitable for
Berber occupation. Here, between Tin Abunda on the
west and El-Abiad on the east—a distance of about 100 miles
—the expedition identified many thousands of ancient
eraves distributed in groups along the plain and foot-hills.
Other evidences of settlement in the area, though less
abundant superficially, are not absent. The seasonal rains
of the hammada were collected and controlled on a con-
siderable scale by embanked canals (pl. XVIa, p. 101) and
subterranean galleries which presumably represent a skill
imported from within the Roman provinces of the north
(see p. 96). Actual habitations are at present little known,
but a stone-walled ‘promontory-fort’ commands the wadi
half a dozen miles south-west of Germa, and a few simple
square-roomed dwellings are recorded. These various
structures are ascribed tentatively to the time of the Roman
Empire, and a house partially excavated at Germa itself is
more certainly of 3rd-century date, though with Byzantine
repairs.
Of the graves, a majority are cairns, either of roughly
heaped stones or reveted on a rectangular plan with careful
dry-stone walling, sometimes plastered, and with a crudely
domed or pyramidal top. The burial chamber within had
1 Monumenti Antichi, XLI (Rome, 1951), cols. 151-552. Objects brought
back by the expedition are in the Museo Coloniale in Rome.
 THE SAHARA 99

in some cases been corbel-vaulted with rough rubble-

masonry. In front of the tomb was occasionally a bath-like
offering-table of stone, with a main oblong cavity and
smaller compartments along one of the edges. With these
tables might be associated low stelae, which sometimes
bifurcate into two projections resembling horns. In-
humation was the rule, and the skeleton was flexed with
varying orientation. But in the present context the im-
portant feature of the graves is less their structural form than
the fact that not a few of them include pottery, glass and
lamps of Roman origin (pls. XVIB, p. 101, and XVIII,
p. 109). Here a few examples must suffice.
About three miles east of the Germa mausoleum, of which
something will be said later, is a cemetery containing a notable
tomb of the plastered square or oblong type already men-
tioned. It had been surmounted by a stepped pyramid,
preserved to a maximum height of 44 ft., with two bicorn
stelae on the first step of the eastern end. Within was a
rubble-faced burial chamber of trapezoidal plan, 8 x 10}x
10x74 ft., with a depth of 6 ft. The tomb had been broken
into previously, the bones scattered and the grave-goods
smashed. Nevertheless, the following fragments indicated
the former richness of the equipment: two Roman clay
lamps of late 3rd- or early 4th-century date; a cylindrical
glass bottle with incised square panels, probably of the 3rd
century; a green glass cup decorated with simple blue spots
alternating with triangular groups of smaller blue spots, a
type also of the late 3rd or early 4th century; a cylindrical
beaker of similar fabric; parts of a green glass vessel bearing
Greek letters; other scraps of glass, a stone roundel or
gaming-counter, and two pottery cups with incised recti-
linear patterns picked out in yellow, orange and white
colours. The pottery cups are non-Roman, but whether of
local manufacture is less certain. The Greek lettering
suggests without proving an Alexandrian origin for at least
one of the glass vessels. The quantity of these in a single
grave is remarkable, and indicates a fairly close contact
100 ROME BEYOND THE IMPERIAL FRONTIERS

between the Fezzan and the Mediterranean about A.D. 300

or a little later.
A similar but rather larger tomb nearby was of the same
general date, and produced further examples of the non-
Roman polychrome ware, together with beads and a bone
comb, and with lamps, amphorae, a glass beaker and other
objects from the Mediterranean world. A third tomb of the
group was even more amply furnished and was probably
that ofa lady ofrank. It was of plastered, dry-built masonry
and retained two steps of a pyramid surmounting a rec-
tangular base about 244 x 224 ft. on plan. Immediately east
of the tomb was a stone offering table of the kind described
above, with the remains of two bicorn stelae. Amongst the
contents were five red-glazed plates (possibly of the and
century A.D.), two amphorae, and several plates and cups of
cut or moulded glass, one bearing a fragmentary Greek
inscription restored on analogy as KA[AWZ TIE], “Drink
well’ or “Your good health’. More notable are fragments of
woollen cloth which possibly represent a purse, and lay on
a stone plate in the south-west angle of the tomb. The
cloth had been variegated with geometric patterns in
yellow, red and blue colours on a purple ground, and was
clearly an imported fabric of some value; it may be recalled
that dye factories along the African coast are mentioned by
Strabo (17. 3. 18) and were doubtless an inheritance from the
Carthaginians. The tomb also preserved by chance in its
plaster the matrices of two hammers with stone heads bound
on to hafts by thongs. Altogether, the equipment was of
mixed character and date, but would appear to have been
assembled early in the 4th century.
In the immediate vicinity of the Germa mausoleum (see
below), extensive cemeteries of somewhat poorer type, with
roughly stone-lined pit-graves, produced other imports,
including amphorae, terra sigillata of the 1st century, glass,
and clay lamps. Two of the lamps bear the maker’s name
IUNI ALEXI (Junius Alexius), whose wares are widely distri-
buted in Italy and elsewhere and are predominantly of the
 XV

A. Green glass beakers from Orslev Underskov, Zealand (left),

and Killerup, Fyen (right). 4 (See p. 86)

g. Green glass drinking-horn, from Lerkenfeldt, Denmark. 3

(See p. 86)
 XVI

A. Ancient irrigation-canal, Fezzan

(See p. 98)

B. Pillared glass bowl frem the Fezzan. 4

(See p. 99)
 THE SAHARA IOI

2nd century. Another, on the other hand, is of a type

common to the 3rd and 4th centuries. Such products occur
in sufficient quantity to minimize the survival factor, and it
may be affirmed that wares from the Roman world were
freely reaching the Germa region from the end of the 1st to
the 4th century a.D. or later, with emphasis perhaps on the
3rd century. Similar contacts, if (so far as is known) in lesser
quantity, have been detected elsewhere along the line of the
wadi for many miles east of Germa; and far away at Gat,
over 200 miles to the south-west, a cemetery of something
like 100 ruined cairns has yielded to cursory examination
several glass vessels to which a 3rd-century date may be
ascribed: notably, an exceptional polychrome beaker decor-
ated with friezes and panels representing a vine tendril, a
wreath, and birds and baskets, probably of East Mediter-
ranean (Alexandrian or Syrian) workmanship.
But at least as remarkable as these fragile imports in the
heart of the Fezzan is the famous mausoleum of Germa
already mentioned, by far the most southerly monument of
Roman type in Africa (pl. XVI, p. 108). It stands alone in
the desert, in its trim, classical dignity defiant of the feature-
less barbarism around it. It was first described by an English
traveller in 1826, and, in spite of its remoteness, has often
enough been visited since that date. Time and tomb-
robbers have dealt drastically with the structure and its former
internal arrangements are doubtful, but externally it con-
stituted a tiny shrine with an eastern portico now repre-
sented by fragmentary columns and Ionic capitals which,
with the gable-ends of the roof, lie scattered on the adjacent
desert. The main body of the shrine had at each corner a
sketchy angle-pilaster defined by the base and ‘composite’
capital carved firmly if crudely upon the quoins; it is possible
that the shafts of the pilasters were originally indicated in
plaster. The base of the three-stepped pedestal on which the
building stands is about 1210 ft., and the height to the top
of the cornice is 15 ft. The ashlar blocks are excellently
squared and fitted, and, save for a provincialism in its
H
102 ROME BEYOND THE IMPERIAL FRONTIERS

carving, the monument might have graced one of the great

coastal cities.
The expedition of 1933 dug out the interior to the natural
soil, and found in the course of their work fragments of
several Roman lamps of the 1st or early and century A.D.,
glass of various colours, Italian red-glazed ware of a type
ascribed to the end of the 1st century, and sherds of less
determinate red and white ware, together with rough sherds
and knives of flint and obsidian of ‘neolithic’ aspect though
presumably contemporary with the other objects. The
occurrence of obsidian is of interest, since its nearest source
appears to be in the Mediterranean islands, Pantelleria,
Lipari and Santorino. The architectural character of the
monument, its contents where datable, and the general
period of the adjacent burials are all consistent with a date in
the latter part of the 1st century.
The adjacent burials referred to lie less than 1 yard to the
west of the mausoleum and consist of two graves under a
heap of sand and stones, each grave containing a Roman
amphora of 1st-2nd-century type with burnt bones. These
two cremation-burials in a region in which the rite is other-
wise that of inhumation are significant, particularly in view
of their proximity to the mausoleum; it cannot be doubted
that they, like the monument itself, represent the intrusion of
visitors from the Roman world, where at this time cremation
was the normal mode.
For the rest, description may here be confined to one other
monument. A short distance to the south of the mauso-
leum are the ruins of a rectangular dry-stone building
about 70x40 ft. externally with two compartments entered
through the eastern wall. The eastern compartment was
probably a court; the western contained a dry-built stone
tomb, square on plan with rounded corners and formerly
roofed as a truncated pyramid. Externally against the front
of the tomb were six rough stelae or orthostats about 2 ft.
high, and again in front of them, in the entrance through the
cross-wall, which was of unbaked brick, was a carefully cut
 THE SAHARA 103
oblong stone trough or offering-table nearly 5 ft. in length,
flanked by two circular stone basins, with a third shallow
basin to the east. Burnt bones of sheep and oxen, mixed
with sherds of wheel-turned pottery, were encountered in
great quantity by the excavators, particularly against the
cross-wall, and there were numerous hearths, and at one
point a flint knife. In and around the tomb were groups of
votive pots, mostly amphorae (sometimes incised with
Berber letters) but also sherds of red ware, with a stone
handmill and a pillared blue glass bowl of a well-known
Ist-century type (cf. pl. XVIs, p. ror). These objects
extended in depth through an appreciable accumulation of
soil and were thought to represent a series of successive
deposits and a considerable length of time. Amongst the
surface-finds was a fragment of an Italian sigillata bowl
bearing the head of Diana, and perhaps of late 1st- or
2nd-century date.
This remarkable assemblage, with its distinctive plan and
imported wares, would appear to represent a long-lived
funerary shrine, of a lower technical grade altogether than
the mausoleum, but of a kind derived also from the north
and rooted perhaps in Carthaginian rather than in Roman
tradition. The extent, however, if any, to which the Fezzan
ritual, with its baetyls or stelae and offering-tables, should be
affiliated to Carthaginian practice must remain undecided
until far more is known than at present of the ritual of the
Punic colonies themselves.
So much for the archaeological evidence as it stands at
present from the Fezzan. That it should fit easily or exactly
into the episodic history of that region in the time of the
Empire is almost too much to hope; nevertheless, the salient
historical facts are worth recalling. In 19 B.c. the proconsul
Cornelius Balbus celebrated a triumph after a successful puni-
tive expedition against the Garamantes whose homeland it
was, and their capital Garama is mentioned by Pliny amongst
the cities which he captured. There is a faint echo of a similar
campaign, undated, against the Marmaridae further cast;
104 ROME BEYOND THE IMPERIAL FRONTIERS

and both enterprises may be seen as moments in the age-long

struggle between the settled civilization of the Mediterranean
littoral and the nomads or semi-nomads of the mountain and
the desert. The antagonism, being of deep-rooted social
rather than political origin, was almost incapable of solution,
and, in spite of intermittent negotiation, the restless oppor-
tunism of the tribesmen long continued to ensure their
participation in any trouble that might be afoot. Thus in
A.D. 69 a domestic dispute between the coastal cities of Oca
(Tripoli) and Lepcis Magna led to the entry of the Garamantes
on the part of Oea and the devastation of the countryside
up to the walls of Lepcis. The incident was the occasion of
the second major Roman expedition into the interior, led
this time by the proconsul Valerius Festus, who followed,
we are told, a ‘short route of only four days’ and cannot
therefore, at the best, have penetrated far into the tribal
area. Writing shortly after the Oca incident, however,
Pliny speaks of the Fezzan tribe as ‘subjugated’, although it
would appear that further punitive expeditions were needed
to confirm the hasty retaliation of Festus. At any rate, by
the end of the 1st or the beginning of the 2nd century
opportunity or necessity had evidently swung the Garamantes
towards the established rule, for Ptolemy the geographer
records two exploratory ventures through their territory
(I, 8, 4-5). The date of these ventures is not given, but they
must have occurred between the time of Pliny (died a.p.
79) and that of Ptolemy (about a.p. 150). On the first
occasion, Septimius Flaccus marched with troops to Garama
and proceeded thence southwards for three months into ‘the
midst of the Acthiopians’, a circumstance which suggests
vague inquiry rather than a measured campaign. The
second venture was led by an otherwise unknown Julius
Maternus, under what auspices is not recorded. Maternus
set out from Lepcis Magna and came to Garama, whence,
with the co-operation of the king of the Garamantes, he led
his expedition towards the south for four months against the
‘Aethiopians’ and reached the unidentified country of
 THE SAHARA 105
Agisymba, ‘where the rhinoceri are wont to assemble’.
Both of these enterprises were doubtless designed, at least in
part, to secure information at source regarding commodities
such as ivory, precious woods and Aethiopian slaves, familiar
to the Roman world through middlemen whose interest it
was to preserve trade secrets and sustain prices. Of these
middle men the Garamantes must have been amongst the
chief, and the archaeological evidence summarized above
suggests that they continued to profit from their position for
two centuries or more after the recorded missions.
A review of that evidence in the light of the historical
episodes suggests a certain measure of agreement between the
two. The Roman imports in the homeland of the Gara-
mantes appear to begin at the end of the rst century a.D.—
that is, precisely at the time when a sufficiently durable
rapprochement was first established between the Empire and
the Fezzani nomads. The direct interest of the Rome world
in the desert and beyond, expressed historically by the
journeys of Flaccus and Maternus, is expressed archaeo-
logically by the Germa mausoleum, which presumably
commemorates a Roman agent and his companions or
successors established at the Garamantian capital with the
goodwill of the king. It was thus that at an earlier date
Roman pioneers had settled far beyond the northern frontier
at the capital of the Marcomanni in Bohemia with the good-
will of Maroboduus (pp. 8 and 20). But it may be that
another analogy is in fact more apt to the Fezzan episode.
Within a decade or two of the erection of the Germa
monument, Domitian was attempting to quicten the formid-
able Decebalus, king of the Dacians in central Europe, by
sending him money ‘and artisans of every trade, both
peaceful and warlike’ (p. 10). Is a similar policy manifest at
Germa, not only in the presence of this surprising monument,
but also and above all in the impressive irrigation systems of
the hammada by the Wadi el-Agial? The date of these is
1 For an interedting discussion of these expeditions in their geographical
context, see F, Rennell Rodd, People of the Veil (London, 1926), pp. 322ff.
ee
%
106 ROME BEYOND THE IMPERIAL FRONTIERS

unknown, but it is difficult to dissociate them from the phase

of Roman contact. A primary object of the Roman
provinces along the Mediterranean littoral must have been to
anchor the predatory nomads, who, on their swift camels,
were a proved menace from the deserts of the south. How
better to anchor them than by turning them into food-
producers, by teaching them to till their own deserts? To
the military pacification momentarily effected under the
Flavian emperors, a systematic instruction in agricultural
engineering was a proper long-term sequel. It is perhaps as
likely therefore that the Germa mausoleum represents a
Romano-Libyan agricultural technician or adviser who died
on a mission to the court of the Garamantes, as that it
represents an enterprising commercial agent or a member of
what we to-day would call the consular service.
Be that as it may, the evidences of trade are likewise
sufficiently apparent in the archaeological record which has
een summarized above. The question arises, What was its
material basis? What did the Garamantian caravans deliver
to the coastal cities in exchange for the wine, the pottery, the
cloth, and above all the varied glassware which these remote
people prized and buried with their dead? Gold-dust,
ostrich eggs and feathers, ivory, precious stones and woods,
animals, slaves have been lightly mentioned, but it must be
admitted that specific historical evidence is of the slightest.
Strabo at the end of the rst century B.c. tells us that ‘the
land . . . of the Garamantes . . . is the land whence the
Carthaginian Stones [i.e. carbuncles] are brought’ (17. 3. 18),
and Pliny nearly a century later has a similar story (V, 34).
Elephants from the African jungle were familiar, in art if not
in life, to every Roman schoolboy.t_ A component of glass,
natron or carbonate of soda, is said to occur in the Fezzan and
may have been exported thence in exchange for the finished
products. But it is best to admit that, like the equivalent
problem in Free Germany, the economic problem of the
1 See generally S. Aurigemma, ‘L’elefante di Leptis Magna’, Rivista Africa
Italiana, Vl (1940), 67ff.
 THE SAHARA 107

penetration of Roman things into desert Africa is only

partially solved.
Of the routes by which the trade was operated, and indeed
of the routes followed by the Roman military expeditions
into the Fezzan, it can only be said that there are three
principal possibilities with a considerable choice of variants.
A westerly track led from Sabratha on the coast south-west-
wards to Ghadames, the ancient Cydamus, near the Tunisian
border and thence south-eastwards to the heart of the
Fezzan. An easterly track passed from Lepcis Magna
through Beni Ulid and Bu Ngem (perhaps the ancient
Vanias) to Hun, at the foot of the Gebel es-Soda, which has
been identified with the ancient Mons Ater or Niger, and
thence southwards to the Fezzan. Between these was a more
direct route from Oca (Tripoli) via Mizda and Gheria.
This central track, as is shown by milestones along it, was
much used in the 3rd century A.D. if not earlier. In the
absence, however, of relevant archaeological evidence,
attempts to allot the recorded expeditions to individual
routes are of no value.
ii. TIN HINAN

So much for the Fezzan. Elsewhere the desert offers yet

another problem of a more individual and elusive kind.
Nearly 1,000 miles south of Algiers the mountainous back-
bone of the Sahara ends in the volcanic massif of the Hoggar,
to which the cliché ‘Mountains of the Moon’ might aptly be
attached. The emphatic, jagged peaks and craters in their
barren setting at the same time repel and fascinate: forming a
theatrical background to a strange, lonely tomb which will
remain a mystery unresolved.
At the western foot of the massif, not far from the oasis
of Abalessa, a rounded hill rises some 125 ft. above the
junction of two wadis. Its slopes are strewn with rocky
debris from the remarkable ruin which occupies its summit.
The ruin, pear-shaped on plan with a major axis of about 88
ft., resembles nothing so much as a stone-walled Scottish dun.
108 ROME BEYOND THE IMPERIAL FRONTIERS

(fig. 14). Its dry-built curtain wall varies in thickness from 44

to 11} ft., and it contains 11 rooms or courts, roughly oblong
save where they conform with the rounded curtain, and
standing to a height of about 6 ft. In the eastern wall is a

THE MONUMENT OF TIN HINAN

IN THE HOGGAR, NORTH AFRICA

1S Metres

Fig. 14 Plan of the ‘palace’ and tomb of Tin

Hinan. (After M. Reygasse)

simple entrance 5 ft. wide. Large iron nails and fragments

of charred wood presumably represent the roofing.
The structure was clearly a small fort or fortified residence
—palace would be a grandiose term. It has impressed itself
upon the minds of the local Tuaregs, and their traditional
story has been recorded. Long ago, we are told, there came
to the Hoggar from the remote Tafilalet a lady of noble
 XVII

Mausoleum at Germa, Fezzan. Height rs feet

(See p. 101)
 (oor “d ag) 8 ynoqy
AjI
-A

(ort -d
Qn
dur}
wouI le
3aL S

od
UWLO] ‘uezzaj oy} UT [eng v wor vioydup “Vv
ue

uly, fo JOLIE
d

af
{1

ur oy}
‘ueUTFy
>
XVIII
 THE SAHARA 109

birth named Tin Hinan, accompanied by a faithful woman

attendant, Takamat, and a number of slaves. The reason
for her journey is not stated, and indeed does not greatly
matter. She rode on a superb white camel, and brought
with her from the land of the Berbers several loads of dates
and millet. But the way to the Hoggar was long, and
hunger began to menace the caravan; until one day the
resourceful Takamat saw a number of anthills and, dis-
mounting, proceeded to rob the ants of their laboriously
stored grain. Thus fortified, they eventually reached the
Hoggar, and in some fashion not elaborated Tin Hinan there
became the ancestress of the Tuareg nobility. The structure
now in question is her monument.
Excavations in 1926 and again in 1933 produced astonish-
ing results. Room no. 1, in the south-western corner of
the complex, had been turned into a tomb-chamber or,
rather, tomb-antechamber. Its two doorways had been
blocked up, and in the floor a number of large flat slabs
covered a basement-cell about 7} x44 ft. on plan and 5 ft.
deep which was the actual grave. In it lay the skeleton of the
traditional first queen of the Hoggar, Tin Hinan. The bones,
now with the other relics from the site in the Musée du
Bardo at Algiers, have been pronounced to be those of a
white woman ‘strongly recalling the Egyptian type of the
Pharaonic monuments, the type of the upper classes, charac-
terized by height, slenderness, wide shoulders, narrow hips
and slim legs’. Of that description, all that can be said is that
it gallantly meets the demands of the legend.
No less can be said of the other contents of the grave. At
the foot of the walls were fragments of matting, in the centre
the remains of a bed or couch of carved wood on which lay
the skeleton, slightly flexed and facing east, with powdery
‘fragments of red leather’ which may indicate that the body
1 For this and an alternative legend, and for the monument itself, see E. F.
Gautier and M. Reygasse, ‘Le monument de Tin-Hinan’, Annales de l’Académie
des Sciences coloniales, VII (Paris, 1934), and M. Reygasse, Monuments funéraires
préislamiques de I’Afrique du Nord (Paris, 1950), S8ff.
110 ROME BEYOND THE IMPERIAL FRONTIERS

had been rouged in accordance with a widespread custom.

On the right arm were eight bead bracelets of silver, on the
left seven of gold; at the left shoulder a small stone cup con-
tained ochre, and by it a little bundle of plants; on the breast
lay a gold pendant and beads, white and red, of agate,
amazonite, calcedony, carnelian and glass. Behind the head
and round the neck and shoulders were about 100 barrel-
shaped silver beads, and at the left hip were some 30 other
stone beads of various colours. The right foot was sur-
rounded by beads of antimony, and beads of stone and metal
were beside the left foot. At the left shoulder, cloth which
fell into dust on discovery had perhaps been pinned by two
iron pins. Nearby were baskets containing dates and grain,
together with two wooden milk-bowls, traces of a coin
of Constantine the Great, a gold ring, two small gold balls,
fragments of glass with cut geometric pattern of 3rd-4th-
century type, and a pendant of polished gypsum representing
in stylized form a grotesquely developed female, obviously
wornasafertility-charm. In this mixed assemblage, the Con-
stantinian coin and the glass indicate a 4th-century date and
some now undefinable contact with the Mediterranean.
The beads are of a rudimentary kind which occurs anywhere
from the Ivory Coast to Carthage. The charm is of a
widespread class without close specific analogy. The legend
of Tin Hinan does not yield to archaeology.
Elsewhere in the little fortress the excavators found a
pottery lamp of 3rd-century Roman type (pl. XVIIIb,
p. 109), fragments of another, a bracelet of twisted iron and
another of bronze, more beads, barbed arrowheads of
iron and an iron knife. On some of the wall-stones are
fragments of undeciphered Tifinagh inscriptions, in part
mutilated by the cutting of the rough masonry and there-
fore earlier than the building; and at one point a camel
and another animal are scratched on an interior surface.
Round about are several small tombs of the so-called
1The term Tifinagh is used for these characters by the Tuareg, who are
themselves unable to read them.
 THE SAHARA III

‘chouchet’ type—that is to say, a shallow pit-grave sur-

rounded by a circular dry-built wall which derives its name
from a supposed resemblance to the chechia or Zouave cap.
They contain inhumations, but no grave-goods, and merely
testify to the sanctity of the site.
In the absence of comparable relics within hundreds of
miles of Tin Hinan, it is impossible at present to put her
fortlet and tomb into a wider context. The Fezzan is the
nearest known spot whence Roman goods such as the lamp
and glass could have been obtained, but reason—if such be
expected for an isolated occurrence—is lacking. To postulate
that the structure may have been a rather superior block-
house upon a suppositious trade-route is a mere guess. We
may be content to leave Tin Hinan in geographical suspense
midway between the Roman provinces and Timbuctoo,
and to think of her, perhaps, as a Lady Hester Stanhope of
another age.
 VIIl* HAST APRICS

[: the latter half of the 1st century A.D. an anonymous

Roman subject from Egypt, possibly from Berenice on
the Red Sea coast, sailed the Red Sea and the Indian
Ocean in merchant ships and, for the instruction of his
kind, set down in unscholastic Greek a factual and remarkable
account of the busy trafficking of those parts. His book, the
Periplus of the Erythraean Sea, is a social and geographical
landmark of the first order; I should describe it, indeed, as
one of the most fascinating books that have come down to us
from antiquity. It will be cited extensively in Chapter
IX. Here it is relevant to note the author’s knowledge—
somewhat vaguer, be it admitted, than usual—of the east
coast of Africa, from the mouth of the Red Sea to Zanzibar,
and to remark how little his lead has been followed up by
archaeological investigation. Turning southward from the
Cape of Spices (Cape Guardafui), he takes us in succession
to the seaport of Opone (Ras Hafun?), the anchorages
of Sarapion and Nikon, the Pyralaae islands, the island
of Menouthias (Zanzibar?), and finally the market-port of
Rhapta. (The knowledge of Ptolemy in the 2nd century
extended over 400 miles further south, probably to Cape
Delgado.) The name Rhapta the Periplus derives from the
rhapta ploiaria, boats sewn together with string, anciently a
widespread mode of ship-construction which still survives,
for example, along the coasts of southern India. The
location of the town is uncertain, but is presumably within
a reasonable range of Dar-es-Salaam. The interesting
information is added that the native ruler of the coast
hereabouts had delegated some special measure of authority
to the distant trading-town of Muza, near the mouth of the
Red Sea, and that Muza sent thither ‘many large ships, using
Arab captains and agents who are familiar with the natives and
intermarry with them and know the region and understand
 EAST AFRICA 113
the language’. This was evidently a survival from the
pre-Augustan era, when ocean trade was still largely in the
hands of middlemen, amongst whom Arabs predominated.
The exports from these east African stations to the Roman
world included, we are told, ivory, rhinoceros horn, tortoise-
shell (‘in best demand after that from India’), palm-oil,
cinnamon, frankincense and slaves, ‘which are brought to
Egypt in increasing numbers’. In return the African
markets received ‘lances made at Muza especially for this
trade, hatchets, daggers, awls, and various kinds of glass;
and at some places a little wine and wheat, not for trade but
to secure the good-will of the savages’. The apparatus of
African trade, whether lances or muskets, wine or gin,
altered very little through the ages! Incidentally, the
cinnamon recorded as an African product was probably in
fact supplied by Indian traders. The Periplus affirms that
the African seaports were in contact, not only with Egypt,
but also with India, whence ships regularly brought ‘wheat,
tice, ghee [clarified butter], sesame oil, muslin, girdles and
sugar. Some of the Indian ships discharged their cargoes
actually in the African ports, but others transhipped them at
sea—a practice due perhaps to the jealous exclusiveness of
Arab middlemen, but more probably to the enterprise of
Arab traders attempting to forestall competition ashore.
Material relics of this far-flung east coast commerce are at
present almost completely lacking. Indeed, only at one
point is there a hint of things to come. In the neighbour-
hood of Port Durnford, 250 miles north-east of Mombasa,
Captain C. W. Waywood found in 1912 ‘a walled-in
fortress, enclosing about 5 acres of ground. He caused his
native servants to dig over the top-soil in places and was
rewarded with the discovery of the following copper
coins.’ The list comprises one Ptolemaic (3rd-1st century
B.c.), one each of Nero and Trajan, two of Hadrian, one of
Antoninus Pius, one uncertain but of the 1st-and century
A.D.,and seemingly 79 of the 4th century a.D., amongst which
1H, Mattingly in Numismatic Chronicle, sth series, XII (London, 1932), 175.
114 ROME BEYOND THE IMPERIAL FRONTIERS

coins of Constantine I and II and Constans are mentioned.

The significance of the discovery is not necessarily vitiated
by the addition of six coins of the Mamelukes of Egypt and
seven of Egypt under the Turks (17th and 18th centuries).
The place may well have been a port of call from Egypt at
many different periods, and is certainly worthy of fresh
examination. The Rev. Gervase Mathew, in a letter to the
writer, describes it from personal knowledge as ‘an almost
land-locked harbour on the Indian Ocean, a very suitable
waiting-place for the monsoon to India, and on my reckon-
ing the “Nikon” of the Periplus’. There for the moment this
alluring problem must be left, pending the emergence of a
worthy archaeological successor to the adventurous Greek.
 PART Ill - ASIA

w we

IX EHESPERIPEUS
|: the previous chapter reference was made to the Greek
merchant’s handbook known as the Periplus of the
Erythraean Sea, and an outline of its contents must now be
given as a preface to a consideration of Roman enterprise in
the Indian sub-continent. Strabo, Pliny, Tacitus, Ptolemy
the Geographer, the map known as the Peutinger Table,
and the Ravenna Geographer all add materially to the
definition of the picture, and picturesque sidelights are
thrown from the less factual literature of India itself; but,
were all these auxiliary sources lost to us, the Periplus would
still preserve a clear and comprehensive outline of Rome’s
remarkable commerce with the East.
By the ‘Erythraean’ or ‘Red’ Sea the writer means, not
merely the sea now known by that name, but all the seas
traversed by this Oriental trade: the Indian Ocean, the
Arabian Sea, the Persian Gulf, and even the Bay of Bengal.
His knowledge often has the actuality of personal experience
and, though now and then it fades apparently into hearsay,
its general accuracy is unimpeachable. He begins by
describing the commercial harbours of the Red Sea (in the
modern usage), and notes Myos Hormos and Berenice,
‘both at the boundary of Egypt’ as ports ‘designated’ for some
special function! which is not specified—perhaps as the
authorized channels for certain types of goods. Further
south on the same western coast was the important port of
Adulis, which also had special legal privileges. From Pliny
(N.H. VI, 103) and others we know that Myos Hormos and
Berenice were both linked by organized caravan-routes with
1 See J. A. B. Palmer in The Classical Quarterly n.s. I (1951), 156ff.
116 ROME BEYOND THE IMPERIAL FRONTIERS

Coptos on the Nile and so with the Egyptian markets.

Between Berenice and Adulis, the small harbourless market-
town of Ptolemais provided some sort of outlet for Meroe,
the decayed capital of Nubia; but it was more important
at this time to observe that behind Adulis lay the kingdom
of the Axumites, in what is now Ethiopia or Abyssinia,
established not long before by immigrants who had been
squeezed out of southern Arabia by combined Arab and
Parthian pressure. Now, in alliance with Rome, it became
an entrepét of African and Eastern trade, particularly as a
focus for African ivory, and received a miscellany of imports
which included gold and silver plate for the king and iron
and muslin from India.
On the eastern and more barbarous shore of the Red Sea,
there was a harbour and fort called Leuké Komé (White
Town), whence there was a road to the Nabataean city of
Petra. The interesting information is added that a Roman
centurion was stationed there with an armed force to collect
one-fourth of the merchandise imported. Further south
and more important commercially was the market-town of
Muza, situated not far from the narrow outlet of the Bab
el-Mandeb. The region was under Arab rule (see p. 112),
but Rome had established trading rights at the town,
placating the king with gifts of ‘horses and sumpter-mules,
vessels of gold and polished silver, finely woven clothing and
copper vessels’. The same king controlled the anchorage of
Ocelis, actually within the straits, which according to Pliny’s
informant (N.H. VI, 104) was the most convenient port for
those arriving from India. This information should perhaps
be read in the light of the further statement of the Periplus
in regard to Eudaemon Arabia (Aden). Eudaemon had
‘convenient anchorages and watering-places sweeter and
better than those of Ocelis’, and, “when the voyage was not
yet made direct from India to Egypt, and when they did not
dare to sail from Egypt to the ports across this ocean, all came
together at this place. It received the cargoes from both
countries, just as Alexandria now receives the things brought
 THE PERIPLUS D7

both from abroad and from Egypt’. Recently, however,

Eudaemon had been destroyed in circumstances that are not
now clear, though the removal of the entrepdt to Ocelis
points suspiciously to the king of Muza as a party to the fact.
Behind the episode emerges the significant circumstance that,
whether in the Gulf of Aden or in the actual entry to the Red
Sea, hereabouts was a traditional barrier where long-range
traffic was withheld from its ultimate Red Sea markets and
Arab middlemen took charge. The establishment of Roman
privileges at Muza marks an attempt to reduce the costly
impact of these middlemen, who presumably still secured
their pickings even though (as the Periplus states) direct
voyages were now undertaken between India and Egypt.
Along the incense-bearing shores of southern Arabia we
need not follow our geographer in detail. Thinking back to
the rich Roman wares from the undocumented regions of
Free Germany, we may again note that the king of Sabratha
in the Hadramaut received ‘wrought gold and silver plate,
also horses, images and thin clothing of fine quality’, either
as propitiatory gifts or as selected imports. Further, we are
told that an unattractive but convenient island, Dioscoride
(the modern Socotra), off Cape Gardafui had become a
trading-post manned by a few hardy agents—Arabs, Indians
and Greeks. Traders from western and southern India
brought to this island ‘rice and wheat and Indian cloth, and a
few female slaves’, and took away in exchange ‘a great
quantity of tortoise-shell’, To the east along the Arabian
coast, Moscha (near the modern Taka) was a specially
privileged port of call much used in the Indian trade; ‘ships
returning from Damirica [South India] and Barygaza
[Broach in Gujarat], if the season is late, winter there and
trade with the king’s officers, exchanging their cloth and
wheat and sesame oil for frankincense’.
Thereafter our merchant turns northward up the Persian
Gulf, at the head of which was a treaty-port called Apologos,
later known as Obollah, near Charax Spasini and the mouths
of the Euphrates. His information in regard to this region
I
118 ROME BEYOND THE IMPERIAL FRONTIERS

is sketchy and perhaps second-hand, but at such brief times

as the Parthian rulers were not at loggerheads with Rome the
port must have been an important meeting-place of sea and
land traffic. To it and to ‘another market-town of Persia
called Ommana’, apparently on the Arabian shore of the
Gulf, large vessels were sent from western India with cargoes
of copper, sandalwood, teak, blackwood and ebony, in
return for local pearls (‘inferior to those of India’), purple dye,
local clothing, wine, dates, gold and slaves.
Of all this Arabian trade, based primarily on the export
of incense and secondarily on the transmission of goods to
and from Egypt and the Orient, scarcely any material
evidence of the Imperial period has yet been brought to
light along its recorded routes. Little search has indeed
been made for such evidence, though regret for this omission
is tempered by the general clarity of the written record.
No doubt the careful excavation of a site such as that of
Adulis would amplify the story in rewarding fashion, and
would in particular give a new precision to the Indian
contribution. To that contribution the second half of the
Greek ‘handbook’ is devoted (fig. 15).
The Periplus first strikes the Indo-Pakistan sub-continent
at the shallow, marshy delta of the river Indus, here trans-
cribed as ‘Sinthus’. The region was at the time occupied by
Scythians under Parthian rule, presumably from the Parthian
metropolis of Taxila in the Punjab. At the mouth of the
Indus was the market-town of Barbaricum, and inland
behind it was the Scythian capital of Minnagara. The site of
neither place is known with certainty to-day, though the
guess that Hyderabad in Sind, at the head of the Indus
delta, may represent alike the ancient Minnagara and the
Patala of Alexander the Great is geographically reasonable.
The imports and exports through Barbaricum form an
interesting list. From the west came ‘a great deal of thin
clothing, and a little spurious; figured linens, topaz, coral,
storax [an incense], frankincense, vessels of glass, silver and
gold plate, and a little wine’. On the other hand, there were
 */ HYDERABAD
( ? MINNAGARA )

ARABIAN

BAY OF
BENGAL
MASULIPATAM
(?MASALIA )

PONDICHERRY & ARIKAMEDU

FI (PODOUKE EMPORION)
| TRANQUEBAR orn KAVERIPATNAM
(XHABERIS EMPORION )

CAPE COMORIN INDIA

HEIGHTS ABOVE 1500 FEET
SHOWING PORTS & MARKETS
e 100 50 O 100
SCALE OF MILES step ed IN THE 1°* CENTURY A.D.
120 ROME BEYOND THE IMPERIAL FRONTIERS

exported costus (a culinary spice and a perfume), bdellium

(an aromatic gum), lycium (a dye and a medicine), nard
(a medicinal oil and perfume), turquoise, lapis lazuli, Seric
(central or eastern Asiatic) skins, muslin, silk yarn and indigo.
Several of these commodities, notably the silk yarn, skins
and stones, must have travelled long distances before they
were shipped at the Indus delta. More will be said of this
at a later stage.
From the Indus our merchant works his way down the
west coast of India, noting especially Barygaza, the modern
Broach, on the coast of Gujarat, predecessor of the later
‘factory’ of Surat on the same coast. The adjacent country,
he observes, was particularly fertile, yielding wheat and rice,
sesame oil, clarified butter (ghee), cotton and the Indian
cloths made therefrom of the coarser sorts. In spite of its
difficult seaward approaches the port was of the first im-
portance in the India trade. From the wealthy cities of
central India it was accessible by the arterial valleys of the
Narbada and the Tapti, and routes of no great difficulty
supplied it from Rajputana and the north. Incidentally,
the Periplus mentions the interesting fact that ‘ancient
drachmas bearing inscriptions in Greek letters and the
devices of those who reigned after Alexander’ were still in
circulation there, having drifted down from Bactria and
north-western India as to-day they still drift (in original or
in imitation) into the hands of the Bombay dealers.
Again, the imports and exports are worth noting. The
former included ‘wine, Italian preferred, also Laodicean and
Arabian; copper, tin and lead; coral and topaz; thin clothing
and inferior sorts of all kinds; bright-coloured girdles a cubit
wide; storax, sweet clover, glass, realgar [a medicinal gum],
antimony, gold and silver coin, on which there is a profit
when exchanged for the money of the country; and oint-
ment’. In accordance with a diplomatic usage with which
we are now familiar, “for the king there are brought very
costly vessels of silver, singing boys, beautiful maidens for
the harem, fine wines, thin clothing of the finest weaves, and
 THE: PERIPLUS 121
the choicest ointments’. The exports comprised spikenard
(a Himalayan herb which produced a valued medicinal
ointment), costus, bdellium, ivory, agate and carnelian,
lycium, muslin of all kinds, silk cloth, mallow cloth (of
coarse, purple-dyed cotton), yarn, long pepper and other
things.
South of Barygaza, Calliena (modern Kalyana, near
Bombay) had at one time been a port offering trading
facilities, but in the time of the Periplus its use was being
obstructed by the local ruler. South again, nine other
places are mentioned, together with certain islands which
harboured pirates. More important were Muziris and
Nelcynda, the former of which may probably be identified
with Cranganore in the Cochin backwaters; the latter was not
far distant, but was situated in another Indian kingdom.
These places were in ‘Damirika’, amended from the Limyrike
of the text to conform with the Ravenna Geographer
(Dimirica) and the Peutinger Table (Damirika), which shows
the region in its Segment XII and marks, incidentally, a
‘Templum Augusti’ at Muziris. The Roman temple has
not been found nor have small-scale excavations at Crang-
anore revealed any evidences of Roman occupation here-
abouts, but the area of potentiality 1s very large and much
further search is required.
The Cochin backwaters are to-day and have long been the
home of a remarkable assemblage of races and creeds:
communities of Jews, Syrians, Roman Catholics alongside
Hindus, each group with its characteristic meeting-places,
sometimes of considerable antiquity. The germ of this
cosmopolitanism doubtless goes back to the days of Imperial
trade, and a Roman temple amongst the palms would accord
with the heterogeneous traditions of the scene. The prime
attraction to the western trader was the accessibility of these
sheltered waterways to the pepper which, then as now,
flourished in the hinterland of the moist Malabar coast.
Much could be and indeed has been written of pepper and its
bearing on the ways of man. To leave the Periplus for a
122 ROME BEYOND THE IMPERIAL FRONTIERS

moment, Pliny (N.H. XII, 14) has an eloquent passage

from the Roman standpoint: ‘It is quite surprising that
the use of pepper has come so much into fashion, seeing
that, in other substances which we use, it is sometimes their
sweetness and sometimes their appearance that has attracted
our notice; whereas, pepper has nothing in it that can plead
as a recommendation to either fruit or berry, its only desirable
quality being a certain pungency; and yet it is for this that we
import it all the way from India! Who was the first to
make trial of it as an article of food? And who, I wonder,
was the man that was not content to prepare himself by
hunger only for the satisfaction of a greedy appetite?’ From
the Roman world the taste for pepper spread to barbarian
Europe; it will suffice to recall that Alaric the Goth demanded
3,000 |b. of it in his treaty with the Romans in A.D. 408.
And, to carry the story to a much later date, in 1592 the
pepper carried by a Spanish East Indian carrack captured by
Frobisher’s ships off the Azores was alone worth £102,000.
In retrospect it is scarcely surprising that the filling of the
Pepper Barns beside the Tiber was a primary function
of Roman trade with the Orient.
In addition to pepper, the Periplus includes amongst the
exports from the Malabar coast ‘great quantities of fine pearls,
ivory, silk cloth, spikenard from the Ganges, malabathrum
[an equivalent of cinnamon] from the places in the interior,
transparent stones of all kinds, diamonds and sapphires, and
tortoiseshell. This is a fairly comprehensive list. The
diamonds, if such they really were, probably came from
central India; the pearls coastwise from the neighbourhood of
Cape Comorin; the silk by sea from China. The popularity
of the market was doubtless due in part to the fact that it
could be reached from the Red Sea by a direct trans-oceanic
route of which more will be said, with the added advantage
that the perils of coastal piracy were thereby reduced to a
minimum. The corresponding imports from the west were
‘primarily a great quantity of coin; topaz; thin clothing, not
much; figured linens, antimony, coral, crude glass, copper,
 THE PERIPLUS 123
tin, lead; wine, not much, but as much as at Barygaza;
realgar and orpiment [a yellow sulphide of arsenic for
making yellow paint, from the Persian Gulf]; and wheat
ee for the sailors, for this is not dealt in by the merchants
there’.
South of Muziris and Nelcynda the Periplus is more
summary and confused, but the general sequence of place-
names probably remains correct. After the Cape of
Comari (Cape Comorin) we come to Colchi, famous for its
pearl-fisheries, which were worked by condemned criminals;
and thence, proceeding from south to north, we reach
successively the market-towns of Camara, Podouké and
Sopatma. Of these, Camara is uncertain unless it be a
variant of Ptolemy’s Khabéris emporion at the mouth of
the Khabéros or Kaveri (Cauvery) river, where the present
Tranquebar represents the former Kavéripatnam or Kaverip-
pattinam. Podouké may be equated with Pondicherry or
Pudu-chchére (‘Newtown’), near which an Indo-Roman
emporium has in fact been discovered (p. 145). With less
evidence, Sopatma has been identified with the So-pattinam
of Tamil literature, the modern Markanam, on the coast
between Pondicherry and Madras.
Ceylon is referred to as ‘the island Palaesimundu, called
by the ancients Taprobane’ but was clearly unknown to the
writer, for a reason which will be suggested at a later stage.
Meanwhile, we may follow the Periplus up the east coast
of India, past Masalia, which may be the port of Masulipatam,
where ‘a great quantity of muslin is made’, to the mouth of
the Ganges, ‘and near it the very last land toward the east,
Chrysé’ or Golden, presumably the Malay peninsula, known
to Ptolemy as the Aurea Chersonesus. On the Ganges was
a market-town of the same name, through which were
brought malabathrum, spikenard and pearls, and muslins of
the finest sorts; and gold-mines and gold coinage (presumably
from the Kushana empire further west) are mentioned.
North of Chrysé lay China, the land called This, “difficult
of access’, whence ‘raw silk and silk yarn and silk cloth are
124 ROME BEYOND THE IMPERIAL FRONTIERS

brought on foot through Bactria to Barygaza, and are also

exported to Damirica [South India] by way of the river
Ganges’. The Periplus ends with a picturesque description
of the mongoloid traders who come together ‘every year
on the borders of the land of This, a tribe of men with
short bodies and broad flat faces, and by nature peaceable. . . .
They come with their wives and children, carrying great
packs and plaited baskets of what looks like green grape-
leaves.’ One would like to see in this an early reference to
tea, but the commodity was in fact probably malabathrum
from the Himalaya.
So much for the Periplus. Its information is elaborated
by other classical writers, but stands first for actuality.
Ptolemy, however, compiling his geography about a.p.
150 from a variety of sources including, as he tells us, “those
who have sailed from us to those places and have for a long
time frequented them’, had nearly a century’s further general
knowledge behind him and supplements the story. In
particular, as we shall see, he knew significantly more about
Ceylon than did his predecessor, and his information about
the interior of India, however distorted, was not inconsider-
able. Along the coast he largely duplicates but also extends
the earlier list, and it may be that on one important matter of
administration he combines with Strabo, Pliny and above all
the Periplus to throw a little light.
Round the coast of India and Ceylon, from the Indus to the
Ganges, 16 of the coastal towns are singled out by him as
emporia (e.g. Muziris emporion, Pédouké emporion), which
were presumably trading-ports in some special sense. The
suggestion is that in each maritime district there was, amongst
other places, one which was pre-eminently the commercial
port, with rights or privileges which it is not Ptolemy’s
business to define. This recalls the use of the phrase nomimon
emporion (‘lawful market’) by the Periplus in respect of Adulis,
Muza and Apologos, or of enthesmon emporion (‘privileged
market’) in respect of Calliena, near Bombay. In these
instances there is a clear indication of special legal provision
 THE PERIPLUS 125

for trading; and some variant of this provision is presumably

implied in the word apodedeigmenos (‘designated’), which the
same authority attaches to Myos Hormos on the Red Sea
and Moscha on the Arabian coast, though the word may
rather indicate the canalization of certain types of traffic
through these two places. Two passages in the Periplus seem
to have a bearing on the matter: at Calliena a new raja had
cancelled or at least obstructed rights conferred on traders by
a predecessor, to the extent of arresting Greek ships on
arrival; and at Leuké Komé, an emporium ‘for small
vessels’ on the eastern side of the Red Sea, there was, as we
have seen, actually a military post under a centurion to
ensure that the Arabs respected trading rights and to collect
a duty of 25 per cent. on imports. It is evident, by and
large, that the privileges of Western traders were differently
safeguarded in the various foreign ports; that in one place
(probably exceptional) Roman authority, even a Roman
Customs levy, might be guaranteed by a Roman garrison,
and that in another the trafficking was subject to the good will
of a local ruler, who might be encouraged by diplomatic gifts
(pp. 116, 117) to grant and observe concessions, but at long
range retained the power to discontinue them. It is likely
enough that in the emporia were normally posted permanent
agencies of the Graeco-Roman traders, organized on lines
not unlike those of the Portugese, Dutch, Danish, French or
British ‘factories’ in the India trade of a much later date.
Indeed, it is fair to envisage Indo-European commerce of the
Ist century A.D. pretty closely in terms of that of the 17th
century; that is, it was based on mutual advantage endorsed
by Western prestige and sufficiently regulated to ensure
continuity.
 X ' THE MONSOON
OTH the Periplus and Pliny have something of interest
to say about the systems of navigation upon which the
Imperial commerce with India was based. When he
reaches the Indus delta, the author of the Periplus records that
‘sailors set out thither with the Indian Etesian winds, about
the month of July, that is Epiphi: it is more dangerous then,
but through these winds the voyage is more direct and
sooner completed’. The Etesian winds, as Strabo tells us,
brought the summer rains to India, and were in fact the
south-western monsoon which begins regularly at the end
of June and lasts until September. By keeping it on the
quarter, the sailors from the ports near the mouth of the
Red Sea were able to steer a tolerably straight course across
the approaches to the Persian Gulf, ‘quite away from the
land’, to the Indus and Barygaza. Similarly, those making
for Damirica or South India sailed direct a little south of
east across the Arabian Sea, throwing the ship’s head con-
siderably off the wind and in favourable circumstances
making the voyage to Muziris, as Pliny tells us, in forty days.
‘Hippalus’, adds the Periplus, ‘was the pilot who, by observing
the location of the ports and the condition of the sea, first
discovered how to lay his course straight across the ocean.’
Hippalus is one of the great names in the history of
navigation. Without his discovery, or at least his popular~
ization, of the monsoon as a dependable aid to deep-sea
voyaging, regular trade with India would have been im-
possible. The long coastwise journey was alike excessively
tedious, and fraught with recurrent danger from piracy.
The history of the India trade is therefore very incomplete
without some agreement as to the date of Hippalus, and
agreement has not hitherto been reached in the matter.
1 The principal danger arose (and arises) from following seas, liable to over-
take and swamp small, slow craft.
 THE MONSOON 127

Our only other historical authority is Pliny, whose account

must be summarized and considered (N.H. VI, 26). Pliny
distinguishes four stages in the development of navigation
between the Red Sea and India. At first there was only
the long coastal route taken by the fleet of Alexander; small
vessels sailed from Aden, along the shores of southern Arabia
and Makran, and made the Indus and the west coast of the
sub-continent. “In later times it has been thought that the
safest line is to start from Ras Fartak in Arabia with a west
wind, the local name for which is the Hippalus, and make for
Patale [at the mouth of the Indus], the distance being reckoned
as 1,332 miles. The following generation [i.e. the third
stage of our series] considered it a shorter and safer route to
start from the same cape and steer for the Indian harbour of
Sigerus [probably south of Bombay], and for a long time this
was the course followed, until a merchant discovered a
shorter route [our fourth stage], and the desire for gain
brought India nearer; indeed the voyage is made every year,
with companies of archers on board, because the seas are
very greatly infested by pirates.’
It will be observed that neither the Periplus nor Pliny
suggests a date for Hippalus. Moreover, whilst the Periplus
associates Hippalus vaguely with the use of the monsoon in
general, Pliny attaches his name only to the second of his
four stages—namiely, the straight voyage from the Arabian
coast to the Indus. Both writers imply that Hippalus was a
historical, not a contemporary, figure; Pliny indeed separates
him by an ‘age’ or ‘generation’ from the third stage, which
was in turn a ‘long time’ before the fourth stage. It is clear
that Hippalus lived some very considerable time before the
third quarter of the 1st century A.D., to which the com-
position of the Periplus and the Natural History may both be
assigned.
It is therefore historically unlikely that Pliny’s second
stage, specifically that of Hippalus, was as late as a.D. 40-1,
the date to which E. H. Warmington has attributed it in his
excellent book on The Commerce between the Roman Empire
128 ROME BEYOND THE IMPERIAL FRONTIERS

and India. Other writers, notably W. W. Tarn in The

Greeks in Bactria and India, have put the Hippalus stage as
early as 80 B.c. or even earlier. Without reconsideration
of these views, two pieces of evidence which have emerged
more recently may be brought to bear upon the problem.
The first seems to relate to Pliny’s story of the freedman of
one Annius Plocamus who in the reign of Claudius (so Pliny
implies) farmed the collection of Red Sea taxes. The
freedman, while sailing round Arabia, was carried out of his
course by northern gales which swept him in a fortnight
(sic) from the coast of Persia to that of Ceylon. As Warm-
ington remarks: “Here we have a man who did not know the
use of the monsoon winds in order to reach Ceylon.’ If
from this we take the further, somewhat reckless step of
inferring that during a part of the reign of Claudius the use
of the monsoon had not yet been adequately advertised,
then the late dating of Hippalus might appear to receive
some support. Latterly, however, Mr. David Meredith, in
studying the ancient inscriptions of the Eastern Desert of
Egypt, has drawn attention to an extremely interesting rock
inscription, a graffito duplicated in Latin and Greek, in a
sheltered spot beside the old road from Coptos to Berenice,
at a distance of about 68 miles from Coptos.1. The Latin
version reads:
LYSA P. ANNI PLOCAMI VENI ANNO XXXV
Oh INKOIN| IIAES Go Gc

The meaning is clear enough: the graffito is a casual record of

one Lysas, a slave of Publius Annius Plocamus, who came that
way and presumably sheltered from the midday sun in the
35th year of the Emperor’s reign (Kaisaros is added in the
Greek version). This Emperor can only have been Augustus,
and the date is therefore July sth, a.v. 6. Identity of
this Annius Plocamus with Pliny’s is not proved and that of
the two freedmen is not of course suggested, but the coin-
cidence of the name in so appropriate a geographical setting
1 Journ. ofRoman Studies, XLII (1953), 38.
 THE MONSOON 129

amounts to near-proof in respect of Plocamus, and it would

be wise to consider the date of his errant freedman in Ceylon
as likely to have been appreciably earlier than the reign of
Claudius.
Be that as it may, of far greater account is the discovery of
an Indo-Roman trading-station at Arikamedu, the native
name of a waterside tract within a few hundred yards of the
Bay of Bengal two miles south of Pondicherry (p. 145),
which almost certainly equates etymologically with the
‘Podouké’ of the Periplus and the ‘P6douké emporion’ of
Ptolemy. The small area excavated here in 1945 yielded 31
sherds of Italic (Arretine) ware, and a score or more of
additional fragments have been found before and since that
year. Now, Arretine ware of the kinds now in question
seems to have been made first about 30 B.c.; more certainly,
it went out of manufacture about A.D. 45, and its extensive
export to the far side of India may be pivoted on a central
date within the bracket, i.e. within the first two decades of
the Ist century a.p. But the postulation of a series of
trading-posts of this kind—for Pddouké was, as we have
seen, only one of a series—upon the east coast of India
carries with it the postulation of regular trade with the West,
and this in turn implies monsoon traffic. The “Hippalus’
may now be assumed therefore to have been in full and
undisguised use at the end of the reign of Augustus (died
A.D. 14); and incidentally the assumption gives a new
actuality to the statement of Strabo, writing under Augustus,
that from the Egyptian port of Myos Hormos alone 120
ships left for the East every year.
Whether the ‘Hippalus’ was known, and if so to whom,
before the latter years of Augustus is another matter. It
may well have been a trade secret of the Arab middlemen or
their Indian agents long before it became familiar to the
shipping companies of Rome and Alexandria. The apparent
fact that an Augustan tax-collector’s freedman had not heard
of it does not imply a corresponding ignorance on the part
of the professional sea-carriers of Muza or Barygaza. At
130 ROME BEYOND THE IMPERIAL FRONTIERS

least as early as the 2nd century B.c. Indians occasionally

found their way to Egypt, and a Greek now and then to
India; and in the earlier half of the 1st century B.c. pepper
was already reaching the Mediterranean in some quantity,
probably via Syria.1 None of these various circumstances
proves or disproves awareness of the ‘Hippalus’ by Greeks,
Indians or Arabs before the reign of Augustus. Nevertheless
it cannot be without significance that the first flood of
Western coinage to reach India was Augustan; with it were
a few Republican denarii, mostly worn, but Ptolemaic
issues are unknown there, or practically so, and Strabo
(17. 1. 13) observes that prior to his time ‘not so many as
twenty vessels would dare to traverse the Arabian Gulf far
enough to get a peep outside the straits’. The one certain
point, be it insisted, is that the regularization and develop-
ment of the India trade in the principate of Augustus was to
a large extent rendered possible, if not by Hippalus himself,
at least by the fact that his discovery and its further implica-
tions were now for the first time common knowledge.
1W. W. Tarn, The Greeks in Bactria and India (Cambridge, 1951), pp. 370-1.
XI * FROM THE INDIAN STANDPOINT

] eras though it be of the historical framework of

the classical records, Indian literature and epigraphy
contain numerous references to these or other contacts
with the West. As early as the middle of the 3rd century
B.C. the Buddhist emperor Asoka could write in two of the
famous edicts which he had graven on rocks and pillars
up and down India that he was, potentially at least, in
diplomatic relations with the rulers of the eastern Medi-
terranean—Antiochos Theos, king of Syria and Western
Asia; Ptolemy Philadelphos, king of Egypt; Magas, king of
Cyrene; Antigonos Gonatas, king of Macedonia; and
perhaps Alexander, king of Epirus. This is a formidable
list. Asoka cites it in connection with his proselytizing
activities amongst his ‘neighbours’, but, even in the absence of
archaeological evidence, it would not be unfair to suspect that
this high-level neighbourliness had a material basis in the
form of trade, though material evidence for this is lacking.
A century later, Menander, most brilliant of the Greek kings
of north-western India, became (then or subsequently) the
hero of a didactic historical romance, The Questions of
Milinda, which refers incidentally to maritime trade with
‘Vanga [Bengal], or Takkola, or China, or Sovira, or Surat
[on the Gujarat coast], or Alexandria, or the Coromandel
Coast, or Further India’.1 Where or when the Questions
were first written down is uncertain,? but indications
point to a date for the oldest surviving version at or a
little after the beginning of the Christian era, so that the
reference may reflect in part the organized commerce of
1T. W. Rhys Davids, The Questions of Milinda (Sacred Books of the East
Series, Oxford, 1890), II, 269. See also W. W. Tarn, The Greeks in Bactria
and India (1951), pp. 414ff.
2 For discussion, see Tarn, pp. 416ff.
132 ROME BEYOND THE IMPERIAL FRONTIERS

Augustus rather than the actual circumstances of Menander’s

own time. It at least presents India as the focus of a far-
flung traffic, which is itself a little suggestive of the later
period.
In the story Menander (or Milinda) has a council of 500
‘Yonakas’. The number was doubtless exaggerated, but the
general circumstance is not unlikely. The word “Yonaka’
is apparently a Hellenistic variant of the more usual term
‘Yavana’, and is equivalent to ‘Ionian’, i.e. Greek. The word
reappears in a Prakrit inscription at Nasik near Bombay in a
Buddhist cave which has commonly been dated to the rst
century B.C., but may be half'a century later. Here also and
at similar caves at Junnar and Karli in the same region are
eight other inscriptions recording dedications by donors who
describe themselves as Yavanas. The names of these donors
are Indian, and the description of them as ‘Greek’ presumably
implies that they or their forebears had come from cities in
the former Indo-Greek kingdom further north; the Yonaka
indeed tells us that he was ‘Indragnitta, son of Dhammadeva,
a Yonaka of Demetrias’, a Greek foundation in Sind. In
the present context the significance of these circumstances is
that they presume a certain preparedness on the part of
wealthy Indian merchants of the 1st centuries B.c.-a.D. for
further contact with the western world.
Apart from this apparent use of “Yavana’ and “Yonaka’ in a
secondary sense, the Yavanas of Indian literature are normally
Westerners in the fullest meaning of the term. The Tamil
poems of the so-called ‘Sangam’ age of the earlier centuries
A.D. contain repeated references to them.t Thus, ‘agitatin
the white foam of the Periyaru, the beautifully built ships
of the Yavanas came with gold and returned with pepper,
and Muziris resounded with the noise’. In another poem a
Pandya (South Indian) prince is exhorted to drink the cool
1 See V. Kanakasabhai, The Tamils Eighteen Hundred Years Ago (Madras and
Bangalore, 1904); P. T. Srinivas Iyengar, History of the Tamils from the Earliest
Times to 600 A.D. (Madras, 1929); The Silappadikaram, trans. V. R. Rama-
chandra Dikshitar (Oxford, 1939).
 XIX

Indi an ivory statuette from Pomp ell.

 p[OD
. 1otnD YT, SNIIOI
> 1 91999)
pur O1ONY 10q TWO (MOI

Ivou “ANYIIT,
Ul yD
 FROM THE INDIAN STANDPOINT 133
and fragrant wines brought by the Yavanas in their vessels.
The epic known as The Lay of the Anklet (‘Silappadikaram’)
describes vividly the quarter of the city of Puhar or Kavérip-
pattinam at the mouth of the Kaveri river—almost certainly
Ptolemy’s Khabéris emporion: “The sun shone over the open
terraces, over the warehouses near the harbour and over the
turrets with windows like the eyes of deer. In different
places of Puhar the onlooker’s attention was caught by the
sight of the abodes of Yavanas, whose prosperity never
waned. At the harbour were to be seen sailors from many
lands, but to all appearances they lived as one commun-
ity....’ And Tamil rajas employed bodyguards of western
mercenaries, ‘the valiant-eyed Yavanas whose bodies were
strong and of terrible aspect’ and who, equipped with
‘murderous swords’, were ‘excellent guardians of the gates
of the fort-walls’. In this capacity they are said to have
been employed at Madura. Yavana craftsmen were also
sought after in southern India, especially for the manufacture
of siege-engines, whilst in the north, as a 3rd-century legend
has it, St. Thomas was brought to the court of king Gondo-
pharnes (c. A.D. 19-45) at Taxila for his skill as a builder of
‘pillars, temples and courthouses for kings’. In one way and
another, the Yavana in partibus enjoyed a considerable
prestige whether as trader or as settler.
That this various interchange was not unsought by the
Indians themselves is clear enough, alike by the presence of
Indian sailors and agents along the coasts of Arabia and
Africa, as indicated by the Periplus, and by the repeated
arrival of Indian missions at the court of the Roman emperor.
Augustus declares in his monumental record at Ankara that
these missions came to him ‘frequently’; and, though it is
not always easy to distinguish the various occasions from
one another, there is more or less specific evidence for at
least four of them during his principate. Thus at Antioch
an Indian embassy to Augustus was seen by Nicolaus of
Damascus and subsequently noted by Strabo (15. 1.73), who
tells us that it came from ‘Porus’, a king of kings probably in
K
134 ROME BEYOND THE IMPERIAL FRONTIERS

the Punjab, where the name or title was traditional. Owing

to the hardships of the long journey, only three of the
delegates had survived, but they brought with them a letter
written in Greek on parchment together with gifts, including
a man born without arms, snakes, a large river-turtle, a
partridge ‘bigger than a vulture’, and a sophist from Bargosa
or Barygaza who in fanatical fervour burned himself at
Athens and was commemorated there by ‘the Indian’s tomb’
(Plutarch, Alexander, LXIX). The Greek letter affirmed
that Porus was ‘anxious to be a friend to Caesar, and was
ready, not only to allow him a passage through his country
wherever he wished to go, but also to co-operate with him
in anything that was honourable’-—a sentiment which
suggests the effusive adumbration of a trade agreement.
Strabo (15. 1. 4) also mentions an embassy to Augustus from
a king ‘Pandion’, who probably ruled over Pandya, the
southernmost kingdom of India. Embassies were apparently
received by Augustus in Spain in 26 or 25 B.c. and at Samos in
21 B.C.,! and Horace refers to one or other of these missions
in his Carmen Saeculare. Waguer references (for example, by
the 2nd-century ‘Annaeus Florus’) tell us of Indian embassies
bringing elephants, precious stones and pearls to Augustus;
and we are left in no doubt that in his time there was a new
rapprochement between East and West unparalleled since
the time of Alexander.
Later emperors continued to receive embassies from the
sub-continent. To Claudius, as Pliny tells us (N.H. VI, 84),
came four envoys from the king of Ceylon, ‘banished by
nature beyond the confines of the world’ and previously—
as is clear enough from Pliny’s account—unfamiliar with the
West. The mission is said to have originated from the
curiosity aroused there by the castaway freedman of
Plocamus referred to above (p. 128). It obviously did not
understate the wealth of the island; certainly Roman trade
was shortly afterwards extended to it, and Ptolemy in the
following century was able for the first time to give an
1 For these and other embassies, see Warmington, pp. 35ff.
 FROM THE INDIAN STANDPOINT 135
adequate description of it. Again, in A.D. 107 amongst the
embassies received by Trajan was one from India (Dio,
LXVIU, 15); and we are reminded that in his last years, his
thoughts full of Alexander, the old emperor stood on the
shores of the Persian Gulf and, seeing a ship sailing away to
India, exclaimed: “Above all things would I have passed over
to India, were I still young.’ Throughout the 2nd and early
3rd century Indian envoys continued to arrive from time to
time. Hadrian received envoys from the ‘Bactrian kings’,
who are probably relevant to a part of our story; and to
Antoninus Pius came ambassadors from the ‘Indians,
Bactrians and Hyrcanians’. Even later, when the Eastern
trade was under a cloud, Elagabalus, Aurelian and Con-
stantine the Great are recorded nevertheless to have received
missions. The lively interest of India in the Yavanas and
their markets needs no further advertisement.
Yet, were it not for the impressive witness of history, the
magnitude of this trade would certainly not have been
guessed. Material relics of it in the West are few, very few,
for the good and obvious reason that most of them were of
an impermanent kind. Best of the rare survivors is the well-
known ivory statuette of Lakshmi, the Indian goddess of
good luck and prosperity, found at Pompeii and therefore
brought there prior to a.p. 79 (pl. XIX, p. 132).? It is a
charming minor work of Indian art of a sort which is
commonly, though not compellingly, associated with the
Kushana, capital of Mathura (Muttra), south of Delhi, and
may well have found its way out of India through Barygaza.
A case, too, has been stated more than once for the derivation
of the squatting posture of the Gaulish god Cernunnos from
India, and astonishingly Oriental the posture sometimes is.
But alternative explanations are possible and more likely.*
1See M. P. Charlesworth in Studies in Roman Economic and Social History,
ed. by P. R. Coleman-Norton (Princeton Univ. Press, 1951), p. 140.
2See A. Maiuri, ‘Statuetta eburnea di arte indiana a Pompei’, in Le Arti
(Florence), 1938-9, pp. III-15.
3 For a documented review, see P. F. Bober, ‘Cernunnos: Origin and Trans-
formation of a Celtic Deity’, American Journ. ofArchaeology, LV (1951), 13ff.
136 ROME BEYOND THE IMPERIAL FRONTIERS

The squatting or “Buddhistic’ attitude is indeed a recognized

feature of Ligurian sculpture as early as the 3rd or and century
B.C. (for example, at Roquepertuse and Entremont in the
neighbourhood of Aix-en-Province), and is presumably a
native element in this remarkable Graeco-Gallic amalgam.
1 See F. Benoit in Gallia V (Paris, 1947), 81-97; and Benoit, L’Art primitif
méditerranéan de la Vallée du Rhone: la Sculpture (Paris, 1954).
 All - SOUTH INDIA
HE Periplus indicates dearly enough that the traffic
with India was basically of two kinds. There were
goods which were produced by the sub-continent
itscl£, and there were others which reached the Indian
markets from further afield, particularly from central Asia
and China. Trade of the former category may usefully
be. termed “terminal trade’; that of the latter was ‘transit
trade’. Although at certain ports, such as Barygaza, the
two groups of commodities would converge, they were
essentially separate and may be separately considered.
The trade of central and southern India was mainly
terminal trade. The spices, muslins, pearls and jewel-stones
which constituted the bulk of it came principally from those
regions. On the showing of the Periplus, there may be
added
od a modicum of transit trade, chiefly in the form of silk
which, diverted by Parthian hostility from the more direct
continental routes, sometimes found its way deviously from
China via the east coast of India to the ports on the Malabar
coast and so joined up with the direct monsoon route to the
West. We are reminded that, centuries later, on this same
cosmopolitan coast a colony of Chinese merchants and
craftsrnen established themselves and adequately prospered at
Quilon. But in Roman times it may be suspected that the
China trade normally found its way through more northerly
ports in a manner to be discussed presently.
The importance of the South Indian trade with the West in
the 1st century A.D. has long been underlined by the im-
pressively abundant Roman coinage which has come to light
fortuitously in the peninsula since 1775 (fig. 16). Of 68 finds
which (excluding those from Ceylon) are known from the
whole sub-continent—India and Pakistan together—no fewer
than 57 come from south of the Vindhyas; and, with the
exception of a stray denarius of Tiberius at Taxila in the
 DISTRIBUTION OF
ROMAN COINS
Setgey Nee Gas HOARD
LAID.
ARIK AMEDU
Ley
2nd CENT. A.D. HOARD
Sro-4TH C. A.D. HOARD
*@O®
Wee CANE, /ID), SINGLE
2nd CENT. A.D. SINGLE
3rp-4TH C. A.D. SINGLE
UNIDENTIFIED COINS
 SOUTH INDIA 139

Punjab, all 1st-century Roman coins not associated with

later issues have been found in the south. More precisely
29 finds, distributed through Madras province and the states
of Hyderabad, Mysore, Cochin, Pudukottai and Travancore,
comprise aurei or denarii ranging from Augustus to Trajan
(pl. XX, p. 133). This mass of coinage demands analysis.
The first noteworthy point is that, of the 29 1st-century
finds, at least 20 are known to have constituted hoards,
ranging individually from four or five coins to ‘some
hundreds, if not thousands’. In the circumstances it is
reasonable to suspect that some of the ill-recorded strays
likewise represent hoards.
Secondly, these 1st-century coins are invariably of gold or
silver. There is no authenticated discovery of a Roman
‘brass’ coin of the 1st or 2nd century in India.
Thirdly, coins of Augustus and Tiberius predominate.
After Nero they dwindle markedly.
Fourthly, the gold coins are liable to be either pierced
for suspension or mutilated by a cut across the obverse.
Only one of the very numerous Roman silver coins is
known to have been similarly mutilated.
Fifthly, there is a notable grouping of early coin-finds
across southern India from the western to the eastern coasts.
Of these five points, the first is readily explained in
relation to the second. The significance of hoards in general
has often been discussed, and they are commonly regarded as
savings lost, or rather not recovered, during a period of
disturbance. But it would be easy to overstress this aspect of
the matter, and, whatever the accidents of their final deposi-
tion, it may be suggested that the preponderance of early
Roman hoards in South India first and foremost represents
an essentially economic factor. The great quantity of
gold and silver coins imported by 1st-century Roman trade
into India is noted or implied by more than one classical
writer. The Periplus mentions it in relation to Barygaza
and as first amongst the imports carried by Western
merchants to the Malabar ports; Pliny (VI, ror) remarks
140 ROME BEYOND THE IMPERIAL FRONTIERS

that ‘in no year does India absorb less than fifty million
sesterces’, which presumably represented actual cash rather
than the eumated value of wares; Tiberius may be thought
to have had this extravagant efflux spescially in mind when he
complaine od to the Roman senate of the reckless exportation
of money ‘to foreign nations and even to the enemies of
Rome’ in exchange for gew-gaws (Tacitus, Ann. Il, 53).
But how was all this currency fitted into the alien economies
of recipients far beyond the boundaries of the Roman
Empire?
The clear answer is that, for the most part, it was em-
ployed not as currency but as bullion. In the whole of
peninsular India there was no native currency of gold or
silver to which the Roman coinage could be approximated.
Pausanias (III, 12, 24) in the 2nd century observed that the
Indians exchanged their wares with those of the Greeks
without understanding the use of money. The potin or
lead coinage struck by the Andhra empire of central India
in the first two centuries A.D. is a partial exception to this
rule; and half a dozen scattered denarii of Augustus and
Tiberius found over a period of years in a restricted area of an
Andhra town at Chandravalli, near Chitaldrug in northern
Mysore, if they do not represent a hoard broken anciently,
may here have circulated in some ratio to the native base-
metal issues. Outside the Andhra zone, however, this
possibility does not arise. For the most part, the imported
coins can only have been used as bullion, to be weighed
out in exchange for goods as silver ornaments or scraps may
be weighed out in an Indian bazaar to-day. Their normal
occurrence in ‘hoards’ is a natural corollary. Indeed, the
term ‘hoard’ is in this context largely a misnomer; the so-
called hoard being doubtless a unit of st amped silver or gold
to a total w eight agreed for some specific purchase, or at
least the bulk reserve from which such units could be
detached. Monetary circulation in the ordinary sense was
notin question. The fact that the precious metal was already
subdivided into known and stamped sub-units (coins) would
 Zab
Ss EK |

A. Arikamedu: brick foundations projecting from river-bank. (See p- 145)/

 XXII

Arikamedu: fragments of Roman amphorae. 3

(See p- 140)
Arikamedu: 1-6, fragments of Arretine dishes (1 stamped vise);
7, fragment of Roman lamp. 3i
(Sce p. 149)
 SOUTH INDIA I4I
nevertheless facilitate the trader’s accounting and at the same
time carry prestige with the customer. Pliny (VI, 85) tells a
relevant story of the admiration expressed by the king of
Ceylon for Roman honesty, on the ground that among the
money found on a Roman castaway (Plocamus’s famous
freedman) the denarii were all equal in weight, although the
various figures on them showed that they had been coined
by several emperors.
This respect for Roman integrity was evidently shaken
in India, as in Free Germany (p. 66), by Nero’s debasement of
his silver coinage in A.D. 63. The third of our points above
needs no additional explanation. Just as in Free Germany
the older pre-Neronic issues in pure silver were sought in
preference to the alloyed issues subsequent to the year 63,
so in India silver issues later than Nero hardly occur, whilst
several hoards end with his reign or that of his predecessor.
And to this shaken confidence on the Indian side in the
bullion value of the newer denarii may certainly be added
export restrictions at source, arising from the expressed
anxiety of the Roman treasury at the enormous outflow.
That the India trade continued or even increased after the
time of Nero is clear from Ptolemy and from the recent
archaeological evidence of Arikamedu, which will be con-
sidered below. But it continued mainly in manufactured
goods and raw materials with only a modest reinforcement
from Roman gold and (much later) base metal.
This brings us to our fourth point, the mutilation of the
gold coinage (two examples in pl. XX, p. 133). The facts
are as follows. Apart from occasional coins pierced for use
as ornaments or charms, in at least six of the hoards the gold
coins have, wholly or mostly, been defaced by an incision
across the imperial head; and, since no other feature of the
design, even when representation of the human figure is
involved, has been singled out for this treatment, the purpose
of the defacement was definitely not iconoclasm, but the
cancellation of the piece as a coin-issue. The defaced coins
cover a wide range of time, including issues of Claudius,
142 ROME BEYOND THE IMPERIAL FRONTIERS

Nero, Vespasian, Hadrian and, apparently in One case,

Constantine I.
For this mutilation there is no need to question the
explanation long ago put forward by Sir George Hill that “the
incisions were made in India, in order to put the coins out of
circulation’. But an interesting fact may be added. Except
in one stiipa-deposit, which as a votive offering may be
regarded as a law unto itself}; none of this Roman gold,
whether mutilated or not, is found within the probable
limits of the late 1st- and 2nd-century Ktshana empire,
which included the whole of north-western India and the
west-coast trading-ports of the Indus delta, Gujarat and
Bombay. Within that empire was at this time struck the
only native gold coinage of the period in India, and it was
struck significantly to the Roman standard. In other words,
it was in unconcealed competition with the Roman coinage,
and the suggestion has even been made that it consisted, at any
rate in part, of re-struck Roman aurei. The implication is
that all Roman gold which could be recovered was absorbed
by the Kushana empire and thus regulated or reminted; and
the all-powerful Kushans saw to it that such Roman gold
as was admitted to their border states was removed by
mutilation from possible rivalry as currency, and relegated
to use as bullion or ornament. The fact that most of this
Roman gold is of 1st-century date, whereas the Kushana
empire reached its prime in the 2nd century is readily
explained by Roman export restrictions from the latter part
of the 1st century onwards and the consequence that little
more than gold surviving in trade from the previous period
was now in use.
The fact that a minority of the Roman gold was not
mutilated implies, on this showing, merely the unequal reach
of Kushana interference or a measure of administrative laxity
that requires no explanation in the East. In regard to
imported silver, the question did not arise. As already
remarked, no silver coinage comparable with the imported
1 Num Chron., 3rd Series, XVIII (1898), 320; modified, ib. XIX (1899), 82.
 SOUTH INDIA 143

denarii existed in India in the ist or 2nd century a.D.;

even the Kushans issued none, with a single exception of
Kadphises II in the British Museum and four coins, also
unique, probably of his predecessor Kujiila Kadphises from
Taxilat Thus, unless very doubtfully in the Andhra
kingdom, there was little risk of the intrusion of the denarius
as currency since the country, including the Kushana
empire, was as a whole economically unprepared for it.
We may now turn to our fifth point, geographical distribu-
tion. Be it repeated that a large proportion of the Roman
coins from India has been found in the peninsula, to the south
of the Vindhyas and even to the south of the main Deccan
plateau. Within this vast area, the district of Coimbatore
and its borders, some 250 miles south-west of Madras, have
produced more than the whole of the rest of the sub-con-
tinent put together. From the Coimbatore district alone
there are at least 11 1st-century hoards, running in the
ageregate to many hundreds of gold and silver coins. As
in all such cases, the first step is to refer to the map.
The district of Coimbatore is approached up two major
river-valleys, that of the Cauvery from the east coast and
that of the Ponnani from the west. It lies at the point where
the Eastern Ghats, swinging westwards, merge into the
Western Ghats and conspire with them to leave a transverse
gap, about 20 miles wide and only 1,000 ft. high, between
east and west. To-day through this, the Ponnani or
Palghat or Coimbatore gap, the railway from Madras and
the Carnatic plain penetrates to Calicut and Cochin; and the
traditional use of this route is indicated by the legend which
lands St. Thomas on the Malabar coast near Cranganore and
takes him thence overland to the Madras coast.2 Along this
same route crowd the coins now in question (fig. 17).
Along it at first, we cannot doubt, came the ancient traffic
from Muziris (Cranganore) and Nelcynda to Podouké
emporion and Khabéris emporion (Arikamedu and Tranquebar)
on the opposite coasts, evading the dangerous voyage round
1J. Marshall, Taxila (Cambridge, 1951), I, 68. 2 Warmington, p. 83.
 144 ROME BEYOND THE IMPERIAL FRONTIERS

Cape Comorin. The avoidance of circumpeninsular naviga-

tion was a habit of ancient travelling. The little Cornish
peninsula and the Jutland peninsula were thus short-circuited
at one time or another. So too was peninsular India.
Strabo in the time of Augustus wrote (15. 1. 4) that in his
day only stray individuals had sailed round India to the
Ganges and that they were ‘of no use as regards the history
of the places they had seen’. Pliny, as already related, had no

RY A.D. FROM SOUTH INDIA

O7 PONDICHERRY

6) COIMBATORE

3 ® BAY
OF
BENGAL

CRANGANUR aoe

\)

©@ SINGLE
HoARas COINS o PLACES Rosny
KIL] HE/GH7S ABOVE 2000 FEET 00 SCALE
~ “30” OF
40" MULES
66 80 _100

Fig. 17

recent knowledge of Ceylon until the embassy from the

king of the island came to Rome with scrappy information
in the time of Claudius. The Periplus becomes noticeably
vague when, southbound, it leaves Nelcynda on the Malabar
coast. Indeed Ptolemy, in the middle of the 2nd century
A.D., is our first circumstantial authority for those parts, and
it has been observed that, in writing of Ceylon, he appears to
have been making a display of information that was largely
new to his reader. It is a fair inference that the Roman
 SOUTH INDIA 145

agencies established in the east coast ports under Augustus

and Tiberius were, so far as the Westerners were concerned,
the termini of trans-peninsular routes, and that only towards
the end of the rst century were the western and eastern ports
linked also by regular circumpeninsular traffic. Con-
sistently with this, no coins earlier than Nero and Vespasian
are recorded from Ceylon.
If there are, then, clear geographical reasons for the
short-circuiting of wealth and traffic through Coimbatore, it
is fair to ask why did so appreciable a proportion of it come
to rest there? The question cannot at present be answered.
According to an unconfirmed but plausible Tamil tradition,
the three ancient kingdoms of South India—Chola, Chera
and Pandya—met in the Coimbatore district, and such a
convergence of frontiers, providing alternative escapes, is at
all times a favourite focus of brigandage. The hiding and
loss of some part of this bullion would fit easily enough into
that picture. Other hoards may represent the forgotten
treasury of local prospectors and miners concerned with the
famous beryl mines of the district, or of the owners of the
pepper estates which doubtless spread, as to-day, on and
below the 3,000-ft. contour in the fringes of the district.
There is reason to suppose that search would reveal ample
evidence of occupation hereabouts, and excavation is now the
necessary preliminary to further knowledge. Only at the
eastern end of the zone, on the Coromandel Coast, has serious
digging so far been attempted, and here the results have been
immediate and dramatic.
Two miles south of Pondicherry, the capital of French
India, a former outlet of the Gingee river forms a lagoon
locked to-day by a sand-bar from the Bay of Bengal, from
which it is further sheltered by dunes and coconut palms.
A part of the eastern bank of the lagoon stands some 20 ft.
above the water and from the scarp project the jagged ends
of successive brick buildings to which the mound owes its
being (pls. XXI-XXIV, pp. 140-1). To the villagers the
site is known as Arikamedu; French archaeologists have
146 ROME BEYOND THE IMPERIAL FRONTIERS

preferred to name it from a neighbouring village, Viram-

patnam. The fragments of walling were noted as long ago
as the 18th century, but it was in 1937 that the site first
attracted archaeological attention.
In that year village children brought to a local French
antiquary a number of relics which they had picked up on the
surface, amongst them a gem (now lost) which is reported
to have borne an intaglio portrait of Augustus. Sub-
sequently French and Indian investigators carried out some
useful if summary digging on the site, and amongst the
resultant finds the present writer in 1944 detected sherds of
Italian red-glazed “Arretine’ ware and of amphorae from
the Mediterranean, together with a fragment of a Roman
lamp and a second Graeco-Roman gem, an untrimmed
crystal intaglio representing a cupid and a bird. The
Arretine ware was of the early 1st century A.D. In 1945 a
systematic excavation was carried out for three months by
the Archaeological Survey of India under the writer’s
direction, and the work was resumed for the French Govern-
ment by Mr. J. M. Casal in 1947-8.1 The accumulated
material of Western origin supplements with important
detail the general historical and numismatic evidence given
above, whilst, from the Indian standpoint, it has for the first
time provided a firm and widely applicable datum for the
associated native culture. The excavations have thus com-
bined in a happy manner the interests alike of Western and
of Indian archaeology.
Much of the town has been removed by the river, and its
former landward extent is unknown, but the excavations of
man and nature have identified a nucleus over 400 yards
from north to south along the bank. Under the southern
part of this area have been found the traces of a village
associated with smooth black-and-brown pottery of the
1For the 1945 and earlier work, see Ancient India, No. 2 (Delhi, 1946),
pp. 17ff.; for the 1947-8 work, see J. M. Casal, Fouilles de Virampatnam-
Arikamedu (Paris, Imprimerie Nationale, 1949); also Aspects of Archaeology
(Essays to O. G. S. Crawford) (London, 1951), pp. 354ff.; and Germania 1952,
p- 389.
 SOUTH INDIA 147

distinctive kind used by the builders of megalithic tombs in

South India between c. 200 B.c. and a.p. 50. This village,
like its modern equivalents in the neighbourhood, doubtless
consisted of simple fisher-folk who caught the gullible fish of
the region from the shore or from small outriggers, gathered
the fruits and juices of the palms, cultivated rice-patches, and
lived in a leisurely and unenterprising fashion just above
subsistence level. To it suddenly, from unthought-of lands
5,000 miles away, came strange wines, table-wares far beyond
the local skill, lamps of a strange sort, glass, cut gems. Traders
arrived across-country from the west coast to meet the large
Indian east coast ships of which the Periplus tells us, laden
with gemstones from Ceylon, pearls from Kolchoi (Colchi),
or spices and silks from the Ganges. A small foreign quarter
like that of Puhar (p. 133) came into being, and finally the
village was replaced by a brick-built town, spreading north-
wards to the sea. There is no reasonable doubt that this
new town was the Podouké of the Periplus, the Podouké
emporion of Ptolemy, the Pudu-chcheri or ‘New Town’ of
the Tamils, garbled by Europeans as Puddicherry and Pondi-
cherry. Shifting sands have moved the town a mile or two,
but the name has come down, little changed, through
nineteen centuries.
Of the buildings of this New Town something is now
known, although brick-robbers have upturned much of the
site. At the northern or seaward end, beside the river and at
water-level, was a large, simple brick structure upwards of
150 ft. long, pretty obviously a warehouse. Further south
were courtyards walled with brick and timber, containing
stoutly constructed brick tanks and cisterns, drains, wells and
soak-pits, the last made in a characteristically Indian fashion
of superimposed terracotta rings. South again, a formidable
brick revetment, sloped or battered and surviving to a height
of 6 ft., was traced by Mr. Casal for a distance of 80 yards
eastwards from its broken end on the river bank and was
interpreted as the side of a tank or reservoir, but may rather
have been a defensive revetment. And still further south
148 ROME BEYOND THE IMPERIAL FRONTIERS

scraps of walling have come to light in the much-disturbed

eround. The overall impression is that of storage-accom-
modation towards the mouth of the former estuary, backed
by industrial quarters where, it may be supposed, the
‘Agaritic’ muslins which the Periplus mentions as an export
of the region were made and dyed, and where beads and other
objects of semi-precious stones, which litter the area, were
assembled or worked. The administrative centre, temples
and dwellings of the town have not yet been identified.
Most of the brick buildings explored were constructed
after the red-glazed Arretine ware had ceased to arrive from
Italy, where its manufacture, in the forms and fabric now in
question, came to an end about a.p. 45. On the other hand
the importation of wine-amphorae seems to have preceded
that of the Arretine and certainly continued after the Arretine
had ceased. The main development of the port may in fact
be ascribed approximately to the middle of the 1st century
A.D., although its international usage must have begun nearly
half a century earlier, if the depth of the deposit (9 ft.)
containing amphora sherds under the warehouse be taken
as an index. Moreover, some of the Arretine ware dates
probably from the first quarter of the 1st century A.D. For the
subsequent duration of the town we have at present in-
sufficient evidence. In the industrial area where the tanks or
dye-vats are situated there were two or more phases of
reconstruction, but there was a general continuity in the main
units of the plan, and sherds of Mediterranean amphora
occurred in all strata. In other words, the function and
contacts of the site remained unchanged. To interpret these
factors in terms of time is guesswork; a minimum of a
century might appear to have been required by the renewals,
but there seems to be no compelling reason to allow more
than two centuries, and a terminal date soon after A.D. 200
is suggested.
The quantity of Western pottery in the relatively small
1 This deposit was largely alluvial mud and probably accumulated fairly
rapidly.
 XXV

Bronze statuette of Poseidon from Kolhapur, Western India. 1

(See p. 151}
 XXVI

(a) (8)
Aand Bs. Handle of bronze jug from Akota, Baroda, Western India.
c. Cameo from Karvan, Baroda, Western India. Nearly ‘
(See pp. 151-2)
 SOUTH INDIA 149

areas uncovered is impressive (pls. XXIII and XXIV, pp.

140-1). At least fifty sherds of Arretine, including four potters’
stamps (VIBIE, ITTA, CAMVRI, C.VIBI OF) are recorded, and
others have been found, mostly in the northern (warehouse)
sector where, it may be supposed, the shippers and merchants
were congregated. Of amphora sherds something like
150 are known to have been unearthed over a wider area; and
incidentally an internal incrustation on some of the sherds
has been shown by analysis to contain resin, a traditional
component of certain Greek wines though here (as elsewhere)
perhaps applied deliberately by the potters to render the

Fig. 18 Rouletted dish from Arikamedu. 4}

vessels impermeable. Other wares which may have been
imported from the West include above all an extensive
eroup of flat-bottomed dishes of a hard, metallic pottery,
whitish in section but with a polished slip blackened
internally by inverted firing and with concentric rings of
rouletting on the upper surface of the base (fig. 18). Local
imitations of this ware are distinguished by softer fabric and
coarser rouletting. This ‘rouletted dish’ is widely distributed
in southern and central India and has become what geologists
would call a ‘type fossil’ in the dating of associated Indian
cultures. It will be considered again at a later stage.
L
150 ROME BEYOND THE IMPERIAL FRONTIERS

Roman glass, including a ‘pillared’ bowl of rst-century

type (fig. 19, 2), reached the site in small quantities, and frag-
ments of atleast two characteristic Roman lamps of early date
have been noted. On the other hand, coins of the period,
whether Roman or Indian, are completely absent. This is in
accordance with the postulate, already discussed, that the im-
ported currency was not as a rule circulated in India in the
normal processes of monetary exchange, the country being
economically unprepared for it.
The discovery of Arikamedu is in more than one respect
a landmark in the study of Indo-Roman relations. For the
first time it gives a habitation and a name to one of the emporia
with which the literature and the coinage had in a more
general way familiarized us. The quantity of the Mediter-
ranean material produced by comparatively trifling excava-
tion is a suggestive index of the extent of the international
trade which used the place. This fact, with the early date of
some of the material and the suddenness with which it is super-
imposed upon a purely native and local culture, has substan-
tiated the essentially Augustan organization of regulated
monsoon traffic; whilst the remoteness of the site, on the
further side of India, emphasizes the range of this new organi-
zation, the powerful purpose with which it was reaching out
eastwards to the sources of pearls and silk. The imagination
of the modern enquirer kindles as he lifts from the alluvium
of the Bay of Bengal sherds bearing the names of craftsmen
whose kilns lay on the outskirts of Arezzo. From the woods
of Hertfordshire to the palm-groves of the Coromandel, these
red-glazed cups and dishes symbolize the routine adventures
of tradesmen whose story may be set only a little below that
of king Alexander himself:
Other sites in central and southern India have produced
occasional evidence of direct or indirect contact with Roman
things, but cannot at present be classed with Arikamedu.
Reference has been made to the rouletted dishes which were
first identified and dated there, and to their distinctively
Mediterranean character even when they may in fact be
 SOUTH INDIA 151

local copies. This type, wholly foreign to Indian ceramic

tradition, nevertheless “caught on’ in India and has become
an invaluable index of date on many sites where other time-
evidence is lacking or inadequate. Far from the coast in
northern Mysore, on the great central plateau, it has been
found with, and has helped to date, an Andhra town of the
1st and 2nd century a.p. at Brahmagiri in the Chitaldrug
district; and 45 miles away in the same district it has appeared
at Chandravalli where, appropriately, denarii of Augustus
and Tiberius have likewise been found. The Andhra
towns of Maski and Kondapur, also on the Deccan plateau,
have produced similar sherds; whilst nearer the east coast
at yet another Andhra city, Amaravati, more rouletted
pottery has been gathered on the surface, and the same site is
said long ago to have yielded Roman coins not otherwise
specified. Here, beside the famous stiipa which provided
the sculptures long exhibited on the main staircase of the
British Museum, is a town site which from more than one
standpoint would amply repay excavation. And away to
the north-east, some 700 miles from Arikamedu, near the
sacred temple-city of Bhubaneswar, the excavators of an
ancient walled town known as Sisupalgarh have found
sherds of the same ware. These widespread occurrences
represent altogether but a brief period of fieldwork, and there
is no doubt that our new ‘type fossil’ of the Ist century A.D.
will acquire an increasing importance as time goes by.
Apart from the rouletted ware, red-glazed pottery of non-
Indian, Mediterranean type has recently been found at Nasik,
near Bombay. More manifestly Western in origin are a
bronze statuette of Neptune or Poseidon and a Roman
bronze jug of the middle of the 1st century A.D. from another
site not far from the west coast (pls. XXV and XXVIIA,
pp. 148 and 156). They were found together in a brick
building of Andhra date in an ancient town-mound at
Kolhapur in the southern part of Bombay province, and
are striking evidences of the trade which, as our historical
authorities indicate, used a number of harbours on the
152 ROME BEYOND THE IMPERIAL FRONTIERS

Konkan coast, south of Bombay. When last seen by the

writer in 1948 the jug and statuette were housed in a little
museum at Kolhapur. The statuette is of better quality
than the average commercial product and, perhaps with the
Harpocrates from Taxila (p. 158), is the most noteworthy
work of its kind from the East.
A little further north, at Akota in Baroda State, the
bronze handle of a similar jug bearing a figure of Cupid
(pl. XXVIA, B, p. 149) has been discovered recently during
the excavation, by the Baroda University, of the ancient site
of Ankottaka. It is closely comparable with the handle of
the jug from Hoby (pl. II, p. 21) at the other end of the
Roman map. From Karvan in the same state a cameo of
exuberant sub-classical workmanship has also been recovered
(pl. XXVIc, p. 149), but details are at present lacking.
Also from central India, generally from the territory of
the Andhra empire, come a number of imitations of Roman
coins, made locally as ornaments and mostly pierced or
looped for suspension. They are normally of terracotta and
were doubtless originally gilded. The most remarkable series
was unearthed at the Andhra town of Kondapur in Hydera-
bad State and is preserved in the Hyderabad (Deccan) museum
(pls. XXVIII, XXIX, pp. 156-7). It consists of at least 20
recognizable imitations of aurei or denarii of Tiberius
(d. a.D. 37), with an Indianized version of the head of the
emperor on the obverse and of Livia as Pax on the reverse.
Both sides bear garbled inscriptions. Two similar clay
bullae, one of them again imitating a Tiberius (with Livia-
Pax facing the wrong way, as befits a copy), were found
at the Chandravalli site already mentioned. Another, also
of clay, was dug up on the Kolhapur mound, and yet others
bearing romanized heads come from Ujjain and Sisupalgarh,
and even from Rajghat on the outskirts of Benares, further
north. A stone mould for casting a metal medallion of this
class was found long ago at Besnagar, near Bhilsa in Gwalior
State, and two gold medallions of sub-classical type and
pierced for suspension were recovered at Nagarjunikonda in
 SOUTH INDIA 153

the Guntur district from a Buddhist stiipa ascribed to the

and-3rd centuries a.p. Another pierced gold imitation of an
issue of Antoninus Pius was found with a pierced genuine
aureus of Commodus at Chakerbedha, Bilaspur district, C.P.
This list is not complete but sufficiently indicates the astonish-
ing vogue of Roman coin-types over a large part of central
India in the 1st and 2nd centuries.
At the same time be it emphasized that, for all the vigour
of its impact, the Roman trade implied by these odds and
ends had no appreciable or durable effect upon the cultures
of the peninsula. It left a superficial imprint here and there,
sometimes in remarkably remote places, but nowhere south
of the Vindhyas was that imprint more than a graffito upon
an essentially self-sufficient native fabric. Only when we
move northwards towards the foot-hills of the Himalaya
do we find evidences of a more significant and lasting
penetration of Western ideas, and to this difficult problem we
must now turn.
 XII - PAKISTAN AND AFGHANISTAN

NLIKE the south, India north of the Vindhya Range

produced no great quantity of goods of the kind
demanded by the Mediterranean market. Cotton
was grown there, and the carnelian of Rajpipla was doubtless
in some request then as now. But it is primarily as a channel
for trade in transit from remoter sources that north-western
India comes into the present story. The Periplus, as we have
seen, includes turquoise, lapis lazuli, ‘“Seric’ skins and silk
amongst the exports from the Indus delta, and these com-
modities have nothing in origin to do with the lands south
of the Himalaya. Turquoise is above all a product of
northern Iran; lapis lazuli comes from Badakshan, in the
north-eastern corner of Afghanistan; the Seres, elsewhere
associated with the silk and iron trade, are vaguely central or
eastern Asiatic, and the skins described as Seric are evidently
thought of as derived from High Asia; silk yarn was itself
at this time a Chinese monopoly. For all these wares the
direct passage to the West would have been the so-called
Silk Route, across Iran to Syria. Astride this route, how-
ever, stood the implacable barrier of Parthia. Always in
active rivalry and often at war with the Roman empire,
Parthia blocked the Orient trade by extortionate levy or
actual veto. The well-informed Chinese chronicles record
that the Roman ‘kings always desired to send embassies to
China, but the An-hsi [Parthians] wished to carry on trade
with them in Chinese silks and it is for this reason that they
were cut off from communication’.1 Thus save when on
rare occasions a successful Roman campaign momentarily
reopened the western sector of the Silk Route, this ceased to
Operate as a main artery of trade. It passed indeed out of
Western knowledge; so much so that after Trajan’s advance
1, Hirth, China and the Roman Orient (Leipsic, etc., 1885), p. 42; F. J.
Teggart, Rome and China (Univ. of California, 1939), p. 145.
 PAKISTAN AND AFGHANISTAN 155

to the Tigris early in the 2nd century it was found necessary

to explore the route afresh, if, as seems likely, we should
ascribe to this period the pioneer mission sent by a Mace-
donian merchant, one ‘Maés called also Titianus’ to the
Stone Tower (Tashkurgan, between Yarkand and Badakshan)
and thence to ‘Séra, the metropolis of the Seres’.1 With
the journey of Julianus’s knight to the Baltic (p. 9), this
enterprise ranks amongst the great adventures of Roman
commerce.
The Maés episode alone sufficiently emphasizes the normal
insignificance of the main overland route through western
Asia in Imperial times. Whether or to what extent a route
further north from the Oxus by way of the Caspian to the
Black Sea was found more practicable we do not know.
The Asiatic connections of the Greek cities of South Russia
suggest the possibility of easier access along it, but evidence
is at present lacking. Alexandria at least is unlikely to have
used it, and Alexandria was the focus of the Orient trade.
Nor, in the tight of the Periplus and of archaeology, is there
room for doubt as to the main lines of Alexandrian traffic
with central Asia and China. They lay through West
Pakistan and northern India, where geographical and
political factors combined to facilitate them.
First, geography. On the map, the Himalayan massif and
its extensions give the Indian sub-continent an aspect of
exclusiveness but are in fact penetrable at a number of points.
For example, there are routes from China to the Brahma-
putra in Assam, and it may have been along these routes that
silk sometimes reached the coastal trade of the peninsula
(p. 137). Through Sikkim it is possible to reach India from
the Tibetan plateau. Further west a number of feasible if
arduous routes enter Kashmir from Turkestan; the most
notable of them used the famous Karakoram Pass, a desolate
highway, if such it can be called, from High Asia into trans-
Indus Kashmir. But neither this nor any other of these
northern approaches has played any dominant role in the
1 Ptolemy, I, Il, citing Marinus.
156 ROME BEYOND THE IMPERIAL FRONTIERS

history of commerce. Their historical importance lay

rather in another context altogether, as occasional channels
for the diffusion of Buddhism from India to central Asia and
China. The main entries lay on the north-western frontier
of what is now Pakistan, and converge on the plain of
Peshawar, using at first the line of the Kabul river and, from
the 2nd century a.p., the Khyber Pass.
Through these north-western gaps penetrated one of the
great trade-routes of ancient Asia. It left the east-west
trans-Asian Silk Route at Balkh, the little village which
to-day represents Bactra, Mother of Cities, once a vast
metropolis seven miles in circuit. Balkh lies on the steppe of
Afghan Turkestan some 20 miles south of the Oxus, outside
the former frontier of Parthia, and was a nodal point in
the Asian traffic system. From it the old branch-route strikes
south-eastwards to the Kabul river and India through clefts
in the Hindu-Kush, thus circumventing the more formidable
mountain-barriers further east. However devious, this is in
fact the easiest highway from central Asia to north-western
India. From China, says the Periplus, ‘raw silk and silk
yarn and silk cloth are brought on foot through Bactria to
Barygaza’.
Secondly, the political factors. In the latter part of the
Ist century the route just mentioned was adopted as the axis
of advance by central Asian hordes under the leadership of
one of their septs, the Kushans, who had overrun Bactria
and now overflowed into India. The chronology of the
Kushana empire is still in dispute, but by the beginning of the
2nd century a.D., if not earlier, it extended to the mouths of
the Indus and far into the north Indian plains, thus unifying
the political control of a vast tract of country from the
Hindu-Kush to the sea and so, incidentally, simplifying and
reducing Customs dues. Unlike their Parthian contem-
poraries, the Kushana kings were singularly catholic in their
political and cultural outlook and modelled their gold
coinage (p. 142)—perhaps even some part of their religion—
on that of the Roman West. There can be no doubt that
XXVIII

CIIXU],
ULOT]

-daag)(6$1t
‘snsAuoIC]
Suuasoida
v
aqua

gssnodar

-qviung
I

-
I9Aaqis

nal

‘tIpu]
UIIISO
A

‘
indeyyoy

|
l
UO

yj
1$i
“d aag)
sul
ozuoig
 XXVIII

Terracotta imitations of coins of Tiberius (obverses), from Kondapur,

Hvderabad State. Nearly 4
(See p. 1
cotta imitations of coins of Tiberius (reverses), from Kondapur,
Hyderabad State. Nearly {
(Seer aeIes
 XXX

MITA-ITV
jo yaedJoL vpxe, ‘(dvyiig)
fung
Sv “qv "PIITATIXS
 PAKISTAN AND AFGHANISTAN 157

the Western trade which passed through their territory was

directly fostered by them. To that trade we must now
return.
Be it repeated at the outset that, greatly simplified,
the route followed by this cast Asian trade was as follows
(end map). From the Chinese province of Honan or there-
abouts, tracks converged upon the upper valley of the
Hwang Ho and thence proceeded by alternative courses,
which do not concern us here, to Kashgar and Tashkurgan
on the flank of the Pamirs, and so by the Oxus valley to
Bactra. At Bactra the route divided in the fashion already
indicated. Political conditions permitting, the traveller
could continue westwards by way of Rhagae (Tehran) to
Seleucia or Antioch. On the other hand, he could turn
sharply south-eastwards to Pakistan and India, passing a
number of trade-route cities of which Begram, 45 miles north
of Kabul, and Taxila, on the borders of the Punjab and the
North-West Frontier province of Pakistan, are for our
present purpose the most important. From this sector the
valleys of the Indus and its tributaries offered approaches to
the Arabian Sea; or the journey could be continued, roughly
along the line of the Grand Trunk Road, to the Kushana
capital of Mathura (Muttra), south of Delhi, and thence
south-westwards through Ujjain to the port of Barygaza or
Broach.
Many sites along this immense track-system must be
capable of producing relevant evidence, but only two of
them—those just mentioned—have been sufficiently explored
to throw any considerable light upon the extent and character
of the Western trade. Something must be said of each of
them.
For more than twenty years Taxila has been submitted to
large-scale excavation, and the results have been lavishly
published.1_ Here it will suffice to observe that in the rst
century A.D. it was a walled city, now known as Sirkap,
three-quarters of a mile in length, designed to include an
1J. Marshall, Taxila (Cambridge, 3 vols., 1951).
158 ROME BEYOND THE IMPERIAL FRONTIERS

acropolis on an intrusive, rocky headland, and a ‘lower town’

laid out in regular blocks or insulae on what may fairly be
described as the Hellenistic pattern (pl. XXX, p. 157). It was
at first under Parthian rulers who seem to have owed no
allegiance to the central Parthian power in Iran; but soon
after the middle of the century it passed to the Kushans and
was, perhaps a century later, rebuilt by them a mile away to
the north (now Sirsukh). Of the new foundation we have
little knowledge, but it was the Sirkap city that straddled the
period of the Periplus.
Comparison with the contemporary town of Arikamedu
in the far south is at once inevitable and instructive.
At Arikamedu, it may be recalled that a few months’
excavation revealed a considerable mass of Mediter-
ranean pottery—wine-jars, table-ware, lamps, even gems
and glass. Many years of excavation at Taxila have yielded
only a single wine-jar, three or four fragments of Western
glass (fig. 19, 3), a gem or two, but not even a scrap of
Arretine table-ware, nota single Roman lamp. The contrast
is significant. There was at Taxila, unlike Arikamedu, no
terminal trade in the usual Mediterranean commodities,
wine and crockery. On the other hand, there were other
contacts with the West, of a kind which had in fact, as we
shall see, a far more enduring influence upon Asian thought
or expression. Notable amongst these evidences is a charm-
ing bronze statuette of the child-god Harpocrates, his right
forefinger raised to his lips possibly in a gesture of silence
(frontispiece). On his head are the Egyptian crowns of the
north and the south and, since the cult was centred upon
Alexandria, the figure is likely to have come from there,
in or about the 1st century a.p. A similar statuette will be
noted from the Begram hoard (p. 164).
1From this generalization I exclude a few sherds showing Hellenizing
features in fabric or decoration which probably represent infiltration from
western Asia, but have nothing to do with the Italic imports of the main body
of Augustan and post-Augustan trade. The so-called ‘Greek black ware’ from
Taxila is a widespread North-Indian fabric, most if not all of which was
doubtless made in India itself.
 PAKISTAN AND AFGHANISTAN 159
Other bronze and gold statuettes or reliefs from the site,
such as a Cupid and a ‘Venus’, have been claimed as Graeco-
Roman and owe something to Western influence but are
probably of Indian manufacture. On the other hand, a
silver repoussé emblema or circular decoration, representing

Ml

Fig.19 Roman glass: 1, from Begram, Afganistan; 2, from Arikamedu,

S. India; 3, from Taxila (Sirkap), W. Pakistan. 4

Dionysus or Silenus, is a somewhat crude piece of Western

workmanship (pl. XXVIIB, p. 156); it is probably from the
centre of a dish, but, when lost, was mounted on a silver
stand as a table ornament. With it was found a silver spoon
with rat-tail ridge and cloven-hoof handle of Graeco-Roman
type. Two or three bronze saucepans having tubular handles
with ram’s-head terminals are of similar derivation and are
dated to the 1st century a.D. (again compare Begram, p. 164).
Of the gems, two at least are imports from the Roman world:
a gold-set carnelian bearing figures of Eros and Psyche with
160 ROME BEYOND THE IMPERIAL FRONTIERS

a small child, and a nicolo engraved with a winged cupid

running after a bird and closely comparable with a gem from
Arikamedu (p. 146). A denarius of Tiberius is also recorded,
but otherwise Roman coinage is absent. A famous series of
circular stone toilette trays show Western elements in their
carved figure-groups, but are generally of Oriental work-
manship.
These very various imports and derivatives tell a different
story from the straightforward merchandise of the south.
They hint at the casual acquisition of Western things; at
opportunist purchase or levy from passing caravans carrying
goods which may be broadly described as of the ‘luxury’
class. In view of the huge total extent of the material now
available from Taxila, they do not suggest the purposeful
and regular marketing of Mediterranean wares there. This
conclusion will be discussed further at a later stage. Mean-
while, there is another aspect of the matter which must be
touched upon in summary fashion.
An outstanding feature of the material recovered from
Taxila and its environs is a considerable quantity of sculpture,
in relief and in the round, in stone, clay and, above all, stucco.
If we set aside the innumerable small terracottas as secondary
commercial products, the more monumental residue is still
impressive in bulk, range, and sometimes quality. Most of
it is derived from Buddhist shrines and monasteries; works
of a definitely secular origin scarcely occur. More will be
said at a later stage of the general problem which it raises.
Meanwhile, examples of this art are cited to illustrate a
specific character of some part of it—namely, a recurrent
Western, Graeco-Roman element of a striking and significant
kind.
The first is the stucco head of a smiling child (pl. XXXIa,
p- 164) from the monastery ofJaulian near Taxila. The head
would be in place on any Graeco-Roman site, and has noth-
ing in origin to do with the art of India. A similar comment
is applicable to the stucco head of a fawn or satyr (pl. XX XI,
p- 164) with pointed ears and snub nose, from the apsidal
 PAKISTAN AND AFGHANISTAN 161

temple beside the main street of Sirkap (visible in pl. XXX,

p..157). Again the type is purely classical. A third example
of the kind is the beautiful life-size stucco head of a youth
from the Dhamarajika monastery at Taxila (pl. XXXII,
pp. 164-5). The head is unmistakably reminiscent of 2nd-
century Roman portraits, such as that of the youthful Marcus
Aurelius in the Capitoline Museum at Rome. Other
Westernizing examples might be cited from Taxila, together
with several in which the Western element is combined with
Indian traits. Amongst the latter is a schist frieze from
the Kunala monastery at Sirkap, showing putti and other
figures amidst the serpentine loops of a continuous wreath
with pendant grape-clusters pecked by birds. This classical
motif—a recurrent one in north-western India at the period
—shows, by the fleshiness of the figures and by some of the
postures, the hand of the Indian carver, but is otherwise a
non-Indian conception (pl. XX XIII, pp. 164-5).
Outside Taxila, within the North-West Frontier Province
of what is now West Pakistan and in adjacent Afghanistan,
such Westernizing work is a familiar phenomenon. Thus
from Charsada, on the Peshawar plain, comes a terracotta
version of the Apollo Belvedere (pl. XXXIIB, pp. 164-5), and
from the same region is recorded a stone relief representing
Laoco6n, in Western costume, in the act of prodding the
Trojan Horse, with a very Indian Cassandra in the back-
ground: a strange and revealing mixture of India and the
Mediterranean with a distinctively Western theme (pl.
XXXIV, p. 165). And again a stucco head of a warrior,
also from the Peshawar district, recalls the typical Gaul
of Roman art, whilst at the Buddhist site of Hadda in
Afghanistan a reflection of the ‘Antinous’ type of 2nd-
century Rome is recognizable in the stucco figure of an
effeminate youth, now at the Musée Guimet in Paris.
To these interchanges and resemblances between West
and East we shall return. Before doing so, however, we
must glance at another site, famous in consequence of its
partial excavation by the French Archaeological Mission to
162 ROME BEYOND THE IMPERIAL FRONTIERS

Afghanistan between 1936 and 1942, under the direction

successively of J. Hackin and R. Ghirshman.
Some 45 miles north of Kabul, at the junction of the rivers
Ghorband and Panjshir beneath the towering range of the
Hindu-Kush, are the remains of an ancient town now known
as Begram, but formerly, as it seems, the Kapisi, royal capital
of Kapisa, visited by the Chinese Buddhist pilgrim Hiuan-
tsang in A.D. 644. It lies, as already mentioned, beside the
branch-route from Bactra to India. The site itself falls
topographically into three parts. Dominating the actual
confluence is an area of about 150x100 yards known as the
Bordj-i- Abdullah or alternatively as ‘the Ancient Royal
Town’, fortified by an earthen rampart revetted externally
and internally with unbaked brick. Inadequate excavation
has suggested that this may be the work of Graeco-Bactrian
kings of the and century B.c. About 350 yards to the south,
at the head of a gentler rise, is another oblong fortified area,
280x 80 yards in extent, known as ‘the New Royal Town’.
The slope between the two enclosures also had lateral
defences, but is unexplored.
Present interest centres upon the New Royal Town, where
three main phases of occupation, ranging from the Ist or
and century B.c. to the sth a.p., have been identified. The
first phase dates, perhaps, from the Indo-Greek kings, and to
it belong stalwart fortifications of unbaked brick on stone
footings, with rectangular salients or towers at frequent
intervals and two ditches. On the southern side was a central
gate which has not been planned. Within there is a hint,
but no more, of a rectilinear lay-out as at Taxila (Sirkap).
The phase was succeeded without visible break by Phase 2,
ascribed to the great Kanishka in the second quarter of the
and century. The new régime was marked by fresh build-
ings and repairs to old ones and, in particular, included a

1J. Hackin, Recherches archéologiques 4 Begram (Mémoires de la Délégation

Archéologique Frangaise en Afghanistan, IX, 1939); R. Ghirshman, Bégram,
Recherches archéologiques et historiques sur les Kouchans (Cairo, L’Institut
Frangais d’Archéologie Orientale, 1946).
 PAKISTAN AND AFGHANISTAN 163

‘palace’ of which more will be said presently. Ghirshman

has made a good case for ending the second phase with the
conquest of the territory by the Sasanid king Shapur I about
A.D. 250, basing his conclusion upon evidences of violent
destruction, coins, and a tomb-inscription at Naqsh-i-Rustam
near Persepolis in Iran, which indicates that Shapur extended
his empire to the Indus. After a brief interim, the city was
partially rebuilt and, so far as excavated, appears to have
been destroyed finally by the White Huns after a.p. 450.
The whereabouts of the town visited by Hiuan-tsang two
centuries later has not been ascertained.
Two rooms in the ‘palace’ of the second phase produced a
hoard of Mediterranean and Oriental wares the like of which
has not been seen elsewhere in Asia. At least one of the
doors had been blocked up anciently, and there is no doubt
that the objects had been carefully stored and were not in
actual use at the time of the destruction. Thus in the north-
western quarter of one of the rooms were placed glass
vessels from Syria or Egypt, whilst in the centre and southern
part were coffers and plaques of carved bone and ivory
from India. Elsewhere was a group of bronze bowls from
Western factories, and a cluster of steelyard-weights in the
form of Minerva or Mars (pl. XX
XVa, p. 168). Elsewhere
again were the remains of lacquer bowls from China.
From the ends of the earth—from the West, from the Far
East, from India—exotic things of beauty and worth had
been gathered into the few square yards of these two sealed
rooms, as so much bullion into a bank strong-room.
Associated Kushana coins, including at least one of
Kanishka himself and ending with the last king of the ‘second
Kushana dynasty’, Vasudeva, indicate the general period
covered by the assemblage—namely, the 2nd century a.p.
and the earlier half of the 3rd. The component parts, and in
particular the Western glass which ranges in type from the
(end of the) 1st (fig. 19, 1) to the 3rd centuries, tally. In
other words, the hoard is not an integral deposit, but an
accumulation representing about 150 years. The easiest
164 ROME BEYOND THE IMPERIAL FRONTIERS

explanation is doubtless the correct one. The store was

probably a Customs depot for the receipt of dues in kind
collected by the kings or viceroys of Kapisa from the caravans
which traversed the adjacent highway in the luxury traffic of
Orient and Occident.
Let us for a moment glance at some of those wares which
concern our special study. Statuettes of bronze include a little
Harpocrates similar to that already described from Taxila and,
like it, presumably from Alexandria;1 a Hercules crowned
with the Egyptian calathus and probably of similar origin
(pl. XXXVB, p. 168); a rider in classical garb (pl. XXXVI,
pp. 168-9); and a grotesque ‘philosopher’ of the low-comedy
type of which there are many examples in minor Roman
art. The steelyard-weights in the form of classical busts
have already been mentioned. Bronze feet from stands for
lamps or incense or from caskets represent normal Western
furniture; and bronze bowls identical with some of those
which penetrated into Free Germany occurred here also in
considerable quantity. Alabaster jugs and ‘saucepans’, the
latter with ram’s-head handle, are classical copies of normal
metal forms. The glass includes a great number of familiar
types, from the universal ‘pillared’ bowl of the tst or early
2nd century to a fine bowl of millefiori also of early date,
a vase with a representation of ships and a lighthouse
(tentatively identified by H. Seyrig with the pharos of
Alexandria), painted and cut-glass goblets (pl. XXXVIL,
pp. 168-9), and vessels with semi-detached network sur-
round. But most surprising in this remote spot is a number of
plaster medallions, 6-8 in. in diameter, bearing Mediterranean
reliefs representing Minerva and other classical subjects,
such as the vine harvest or Eros and Psyche, of a kind well
known in the classical world (pl. XXXVIII, p. 169). Such
medallions were used as models by metal-workers, who
transferred them to the silver dishes for which Alexandria
and other Graeco-Roman centres were famous. Alexandria
1P.158. The Begram Harpocrates, now in the Kabul Museum, had had the
right arm wrongly affixed when I saw it in 1946.
XXXI

‘qulung

‘vpixey,

Wory

“Aqes

1 (oot
aay)
-d
jo
ae

OIINIAS
IY

ati
‘qriung
“UTEX
“JuUCJUL CUOTF
T
(oor
2S)“a!
T
ure

pr
jo
LY
oo9onIS

V
 SOX
TI

INNIS ‘qeiung T UJJOOVIIS

peoy WO
poseq oy opjody
Pa IpysAa
‘Dy 1 urCc “‘epesi

C
T,

V
‘

3Y]
oO pr (Nod
rx ea7 I ad
\

UOT
‘CIIXe
 > OO, ONG!

IUOIS OZOTIT ULOTF

T, PEXv
‘qeiung Y Sic rege
098)-d
1SUIT
(191
 (191 «d 209)

MOYS

U0}
Sul
ION, oY} WoIz SsIOPZ urlo
|a}ele
“IOUIAOI
Hale udOsOv
9

ibs
88
N\

l1)
ce

I
5

S
I

J
;
ITUOT
d qs |
XXXIV
 PAKISTAN AND AFGHANISTAN 165

had special cause to use her vast local resources of plaster,

and further reference will be made to the wider significance
of this fact. If in regarding the Begram hoard we ignore
the astonishing beauty of the great collection of carved ivory
from India or the crumbling remains of lacquerwork from
China, and look only at these and other products of Roman
Imperial craftsmanship, the spectacle, whether in the Kabul
Museum or in the Musée Guimet in Paris, where a part of the
collection can be seen, is one which stirs the imagination.
And derelict Begram itself, where so little work has yet been
done though with such dramatic result, remains a challenge
to the explorer.

And now for certain wider issues. If one were asked to

name the most penetrating and enduring impact of the
Roman upon the Eastern world, the answer could scarcely
be in doubt. In the early centuries of the present era the
Buddhist art of north-western India and Afghanistan
absorbed Western modes of expression, varying in kind
and degree from instance to instance and from mind to
critical mind, but constituting in the aggregate one of the
most notorious and intriguing problems of art-history. Nor
is the matter exclusively one of fashions and formulae.
Eastern religious thought was itself undergoing at this time a
process of change which presented significant analogies with
the contemporary developments of Western metaphysics.
The Western imperialistic idea and the Eurasian mysticism
in which it was ultimately enveloped had appreciably close
counterparts in the Orient; so much so that a good deal of
nonsense has been written about the supposed interrelation-
ship of Christianity and Buddhism, and may serve to warn us
against an excessive affiliation of the respective arts. The
main facts, however, are sufficiently clear. East and West
were moving vaguely in the same direction, they had some-
thing of the same spiritual and aesthetic needs. Given the
opportunity, there was every reason for a certain inter-
change or borrowing of material expression. And since
M
166 ROME BEYOND THE IMPERIAL FRONTIERS

1870, when an official of the Punjab service brought to

England from north-western India a small collection of
‘Indo-Scythian’ sculptures, it has been recognized that in and
about the plain of Peshawar—the ancient Gandhara—and
thence far afield in Asia a whole complex of Western
ideas was in fact drawn at this time into the service of
Buddhism.
Gandhara art, and that far wider art-province of which
Gandhara provided a regional facet, does not concern us here
in any detail. Its literature is immense and growing. The
absence of an objective chronology has facilitated an infinite
manipulation of the evidence in accordance with taste
and theory, and until modern methods of excavation are
applied to Buddhist sites far more rigidly than they have
been in the past this source of doubt and disputation will
remain.
Meanwhile, the governing principle may be emphasized
that, with stray exceptions, this Romano-Indian art was
confined to Buddhist patronage. Not only did the Buddhism
of the age and place supply the religious and aesthetic context,
but it also supplied the necessary wealth and backing. From
the arrival of the Kushan dynasty in the latter half of the 1st
century to the devastation of the whole region by the White
Huns in the latter half of the sth, Buddhism went from
strength to strength, Buddhist monasteries multiplied across
the landscape, and within the individual convents sculptured
dedications jostled one another in ever-growing mass about
the central shrine. For an understanding of this tumultuous
manifestation a word or two must be said about the con-
trolling faith.
Buddhism was in origin not a religion, but a philosophy
of life. The Buddha, the Enlightened One, was not a god;
he was an inspired teacher who, about 500 B.c., preached on
the Ganges plains the Middle Path between indulgence and
asceticism and sought an ultimate deliverance from accumu-
lated sin in supreme detachment, nirvana. Later, however,
by a process of evolution natural to a land where the teacher
 PAKISTAN AND AFGHANISTAN 167

has always been revered, the Buddha was increasingly

regarded as a divine Being to whom prayer might be offered.
The earlier type of Buddhism is commonly distinguished as
that of the Hinayana persuasion or the Buddhism of the
Lesser Vehicle, and the later type (which did not wholly
supersede the other) as that of the Mahayana persuasion or the
Greater Vehicle. The latter persuasion reached maturity in
and about the time of King Kanishka, who is now widely
thought to have flourished in the second quarter of the 2nd
century A.D., about the time of Hadrian and Antoninus Pius
in the West.
In artistic expression the outstanding difference between
these two main types of Buddhism was that during the
prevalence of the Hinayana teaching the Buddha himself was
never represented. His presence was symbolized by a chair,
a footprint, an umbrella, a riderless horse. About this
symbol crowd the other participants in the scene, but there
is no central commanding figure. In Mahayana Buddhism,
on the other hand, the figure of the divine Buddha controls
the assembly and is the focus of its composition. Both
iconographically and aesthetically the change was revolu-
tionary.
Artistically, this change found its first full expression
in what is now West Pakistan and Afghanistan. The new
Buddhism, amongst other faiths, received the patronage of
the liberal-minded Kanishka, and the wealth of the Kushan
empire provided a suitable environment for its development.
What was lacking was any comprehensive traditional idiom
in which to express the new observance; and it was here that
Western art, already sufficiently familiar from the Western
luxury trade described above—if not from the actual im-
portation of Western craftsmen (p. 133)—came to the rescue.
In the Roman Imperial West, the figure of the Roman
emperor was established as the dominant feature of an
artistic composition. Now, both this and some of its
accessories and details, with others from Roman heroic and
funerary art, were adapted to the Buddhist problem.
168 ROME BEYOND THE IMPERIAL FRONTIERS

Figures clad in Western clothing, Western types such as

putti or erotes and garlands, satyrs, Apollos, Minervas, even an
occasional scene from Western legend such as that of the
Trojan horse, and Western grouping such as that associated
with the state arrival or departure of the Roman emperor,
found their way into the sculptors’ workshops of Gandhara
and the adjacent region of Afghanistan. There they were to
a greater or less extent transmuted by the Buddhist craftsman
and given a Buddhist context. The fact that the region had
previously been ruled by Indo-Greek kings may have, in
some small degree, prepared the way for this influx. But
it seems certain that, for more than two centuries before the
Mahayana persuasion and its art took shape, the surviving
Hellenism of Bactra had dwindled to vanishing point, and
the most that it can have bequeathed to the subjects of
Kanishka was a faint surviving sympathy for Western things.
The idiom or ‘language’ of the new Buddhist art, in so far as
it Was non-native, was bred from new contacts. It was
essentially a cultural by-product of the Kushana commerce
which brought into and through the kingdom objects and
craftsmanship of the Roman empire.
As examples from Taxila have indicated, this Buddhist art
has survived mainly in sculptural form, partly in stone,
partly in stucco or plaster, and partly in clay. Most of the
stone sculpture is carved in a green schist from the periphery
of the ancient Gandhara. Its comparatively high ‘survival
value’ and portability have tended to concentrate attention
upon it and perhaps to exaggerate its relative importance.
Far more widely spread in space, and probably in time, is the
equivalent sculpture in painted stucco or clay, which is found
not only in Gandhara, but far afield along the arterial routes
that Buddhist monks shared with the traders through the
1 The least Western of these figures is often the much-discussed figure of the
Buddha himself. His monk’s robe occasionally bears a more or less remote
resemblance to the toga, but this resemblance is commonly exaggerated by
modern writers, whilst the head is in most instances completely Oriental. It
is in the details and grouping of subsidiary figures that Western iconography
is more liable to intrude.
ny oZUOIG pavdjoaas ‘JUBIOMwoz 94) UUVIGI “psreoy “a IZUOIG 9IVONAUVIS
JO TOTO
so] OULILIMN
St REETN CAL acta
 XXXVI

Bronze figurine of armed horseman, from the Begram hoard. {4

(See p. 164)
 AAAVITL

Roman glass beaker with painted decoration, from

the Begram hoard. { (Sce p. 164)
 XAAV ITI

Stucco emblemata representing the grape-harvest (below) and

Athena (above) from the Begram hoard. # (Sce p. 164)
 PAKISTAN AND AFGHANISTAN 169

Hindu-Kush and the Turkestans, along the China road. In

Afghanistan the best-known source for this sculpture, mostly
in stucco, but sometimes in stone, is Hadda, near Jalalabad,
where there was a large Buddhist monastic settlement, but
another monastery as far north as Kunduz on the Turkestan
steppe in the same country has produced similar stuccoes,
others are known in central Asia, and in the opposite direction
stucco fragments have been found at Mohenjo-daro in Sind.
It is observable that the stucco-medium, by reason of its easy
manipulation, is associated with a greater range and vividness
of expression than the stonework; and, on the other hand,
that this facility, and the possibility of using moulds, en-
couraged mass production, particularly in the later and more
decadent phases, when the use of the more laborious stone
may have died out.
For the dating of this art there is little evidence. The
earliest undisputed representations of the Buddha are those
on certain gold coins of Kanishka, about or a little before
A.D. 150, and this date accords well with what we know of the
development of the Mahayana observance. A terminal date
for the bulk of the surviving material in and about Gandhara
is more securely fixed by the havoc wrought by the White
Huns there during the half-century following a.p. 450.
At Taxila a considerable quantity of stucco sculpture has
been found in position in monasteries destroyed at that time
and gives us a consistent picture of the condition of the art
about the middle of the sth century. We see that it was then
highly stylized, but still retained traces of its dual (Eastern
and Western) origin. The main brackets for this classicizing
art in its homeland may be provisionally stated therefore as
A.D. 100 and 450, though the tradition is still recognizable
long after this period and in regions as remote as South India
and the fringes of the Gobi Desert.
It remains to consider these various facts and probabilities
in the general context of time and place. How did the
Western elements reach the Gandhara studios? A part of the
answer has been already given by Taxila and Begram.
170 ROME BEYOND THE IMPERIAL FRONTIERS

Behind the Western statuettes, glass and stuccoes from those

sites loom the presence and personality of Alexandria.
Other Western workshops, notably those of Syria, doubt-
less contributed; but the dominance of Alexandria is scarcely
in doubt. Long ago, before most of these discoveries were
made, Rostovtzeff was able to affirm that ‘the active agents
in the exchange of goods between the Roman Empire and
China were the Alexandrian merchants. Without them the
commerce with India would probably not have existed.’
Mote recently a growing knowledge of the art and crafts-
manship of Palmyra has pointed to a certain affinity of
Palmyrene textiles and jewellery with those of India, and
occasionally to comparable traits, which may or may not be
significant, between Palmyrene and Gandhara sculpture.
Such resemblances arise naturally from the known trade-
connections of Palmyra with the Orient, but they merely
supplement the arterial traffic of Alexandria. Above all,
Alexandria was the principal home of stucco sculpture in
the West. Adjoining the city for miles the coastline is
white with the gypsum which is the raw material of stucco.
And when in Ptolemaic and Roman times (from the 3rd
century B.C. forwards) statuary was demanded at Alexandria
in conformity with Greek and Roman taste, the cheap local
stucco was extensively used as a substitute for the relatively
costly white marble which had to be imported from con-
siderable distances. There can be little doubt that it was
the Alexandrian trade with and through the Kushan empire
that carried thither, not merely goods and ideas, but also the
stucco-technique; and that thereafter the use of this mobile
medium spread rapidly with the Buddhist monachism that
travelled northwards and eastwards with the caravans as
far as the border of China.
A rich and dominantly Alexandrian trade and a peculiarly
receptive environment were, then, contributory factors in
1H. Seyrig, ‘Ornamenta Palmyrena Antiquiora’, in Syria, XX1 (1940), 305ff.;
Rostovtzeff, Revue des Arts Asiatiques, VII (1931-32), 209; A. C. Soper, “The
Roman Style in Gandhara’, in American Journ. ofArchaeology, LV (1951), 311.
 PAKISTAN AND AFGHANISTAN 171

the making of Romano-Buddhist art. It may be doubted,

however, whether they operated without a third medium.
Before A.D. 100 India was in intermittent diplomatic contact
with the West—through the various missions to which
reference has been made (p. 133), and, by fair assumption,
through others unrecorded—and was actually employing
Western craftsmen. The ancient story that King Gondo-
phernes purchased St. Thomas as a slave in Jerusalem for
the purpose of building his palace at Taxila may not be true,
but must have been sufficiently possible to achieve accept-
ance; and more historical references to the reputation enjoyed
by Yavana craftsmen employed in India are impressively
numerous. The suggestion that, on analogy, an occasional
Western sculptor was similarly employed in Gandhara for
training and assisting local craftsmen at a time of unpre-
cedented demand is entirely consistent with this trend, and
indeed completes the picture.
In summary, therefore, the basic explanation of Romano-
Buddhist art in north-western India appears to be threefold.
First, there were the new and congenial requirements of
a Mahayana Buddhism backed by powerful patronage.
Secondly, there was close commercial contact in high-grade
wares, including sculpture in bronze and stucco, with the
West and, in particular, with Alexandria. Thirdly, there is
analogy for the direct employment of small numbers of
Western craftsman. in partibus. To these factors may be
added a fourth contributory factor in the incidental im-
portation from Alexandria of a technique, that of stucco
sculpture, which lent itself in a welcome measure to easy
diffusion and mass-production. There the problem in its
present limited setting may be left.
1 See above, p. 133, and A. C. Soper, as cited, p. 305.
 XIV oO TES EA eee AS

} \HE author of the Periplus was aware of lands beyond

India, but knew almost nothing of them. East of the
Ganges lay Chrysé, the Golden, ‘the very last land
toward the east’. Northwards was This, a land ‘hard of
access’, with a great inland city called Thinae, whence silk
was brought on the one hand to Bactria and Barygaza and
on the other hand to the Ganges and South India. A century
later, Ptolemy was able to collect extensive lists of place-
names from the coasts, islands and hinterland of these parts,
but had little knowledge of their interrelationship. He was
indeed misled by a basic conception that the Indian Ocean
was a larger Mediterranean, south-eastern Asia being twisted
southwards and westwards to join the coast of Africa some-
where south of Zanzibar. In the circumstances, it is not
very profitable to spend time on an attempted reconciliation
of his data with fact. His ‘Golden Chersonese’, like the
Chrysé of the Periplus, may be equated with Burma and
Malaya, and his “Great Gulf’ was probably the Gulf of Siam.
In the Golden Chersonese he notes amongst other places two
markets or emporia, Takola and Sabana, of which the former
has been placed tentatively in the neighbourhood of
Rangoon, where there is said to be some evidence of a place
called Takkhala, Takola or Tagala in the middle ages.1
But only archaeological evidence can now give substance to
identifications of this vague kind, and such evidence has
scarcely been sought. That tangible evidence does in fact
await discovery is suggested by the preliminary excavation
of an ancient town-site known as Oc-eo, 15 miles from the
shores of the Gulf of Siam, in Indo-China.
The site of Oc-eo lies by the delta of the Mekong river,
1 On this point, and generally, seeJ.W. McCrindle, ‘Ptolemy’s Geography of
India and Southern Asia’, in The Indian Antiquary XII (Bombay, 1884), 372,
etc.; and McCrindle, Ancient India as described by Ptolemy (Calcutta, 1927).
 THE FAR EAST 173
on the fringe of the province of Long-xuyen, and is a
complex of low mounds rising from the alluvial plain. The
remains are those of a town framed by buried defences which
form an oblong of rather more than 3,000 by 1,500 yards,
with the considerable area of about two square miles. This
area seems to have included timber huts (possibly on piles),
and brick and stone buildings, together with rice-fields.
Excavations were begun here in 1944 by M. L. Malleret on
behalf of the Ecole Francaise d’ Extréme-Orient,! but have
been curtailed by war conditions.
The excavator found remains of rectangular structures
with flat roof-tiles, and more or less completely excavated
two of them. One, of brick, consisted largely of closed cells
which may (I suggest) have constituted the podium of a
temple, comparable with the cellular podia characteristic of
waterlogged sites of 7th century and earlier temples in
Pakistan and northern India.2. The other structure, of brick
and granite, was identified as a mandapa or temple building
with terracotta plaques bearing floral decoration and heads
of monsters of a characteristicaily Indian type. Immediately
flanking it were slight remains of two annexes, in one of
which was part of a stone statuette of the Hindu god Vishnu.
Other objects found during the excavations included relics
of the cults of Siva, Vishnu and Surya, traces of gold-, iron-
and copper-working, many thousands of beads of semi-
precious stones, cameos of carnelian, crystal or jasper bearing
in some cases brief Brahmi inscriptions of the 2nd-sth
centuries A.D., fragments of bronze figurines, amongst
which are noted Buddhist representations, a head and image
in the Gandhara style and a statuette recalling the art of
Amaravati, gold jewellery, and a wide range of terracotta
ornaments. As a whole, the culture is half-native, half-
Indian, and is appropriate to its peninsular location. But
1 Preliminary report in Bulletin de l’Ecole Frangaise d’Extréme-Orient XLV,
fasc. I (Paris, 1951), 75ff., with other references.
2R. E. M. Wheeler, Five Thousand Years of Pakistan (London, 1950), pp.
102-3.
174 ROME BEYOND THE IMPERIAL FRONTIERS

with these Indian elements are a number of objects or

influences from further afield. Thus a blue glass gem bears
in intaglio a bearded head with plaited hair and a Phrygian
cap, probably of Sasanian origin. Other gems are regarded
by the excavator as Roman: one bearing a cock on a chariot
drawn by two mice, others with grylli or composite animals
of a kind familiar in Roman art, others again with busts
of Roman or romanizing type. A few of the beads are
described as “‘pseudo-Roman’, but publication of them is
awaited. No doubt at least attaches to a pierced gold coin of
Antoninus Pius, dated to A.D. 152, and a second, less legible,
of Marcus Aurelius.
All these Western elements should doubtless be ascribed to
infiltration through India rather than to any more direct
contact- with the West. Nevertheless, it is a noteworthy
coincidence that to the time of Marcus Aurelius the Han
Annals attribute the arrival of an embassy from “An-tun’
(Marcus Aurelius Antoninus) at the Chinese court and the
first opening up of intercourse with Ta-ts’in or the Roman
Empire. The entry is as follows: “They [the Romans] make
coins of gold and silver.... They traffic by sea with An-hsi
[Parthia] and T’ien-chu [India], the profit of which trade is
tenfold. They are honest in their transactions, and there are
no double prices.... The budget is based on a well-filled
treasury.... Their kings always desired to send embassies
to China, but the An-hsi [Parthians] wished to carry on trade
with them in Chinese silks, and it is for this reason that they
were cut off from communication. This lasted till the ninth
year of the Yen-hsi period during the Emperor Huan-ti’s
reign [= A.D. 166] when the king of Ta-ts’in, An-tun, sent an
embassy which, from the frontier of Jih-nan [Annam],
offered ivory, rhinoceros horns and tortoiseshell. From that
time dates the intercourse with this country. The Annals
proceed to comment on the absence of jewels from the list of
gifts offered by the ‘embassy’, and the comparative poverty
1P. Hirth, China and the Roman Orient (Leipsic, Munich, Shanghai and
Hongkong, 1885), p. 42.
 THE PAR EAST 175

of the tribute does in fact suggest the opportunism of some

private merchant rather than a considered Imperial mission.
Nevertheless, the episode, which is thoroughly circumstantial
in its attendant details and is unlikely to have been invented,
is a further example of the adventurous spirit in which
Roman trade was conducted in partibus.
For the rest, Roman or sub-Roman wares, particularly
glass, have been found from time to time in China. In
south-eastern Korea two of the mounds which mark the
burials of the kings of the local kingdom of Silla, broadly
bracketed between 100 B.c. and A.D. 600, have produced
glass vessels which may be Roman of the 3rd or 4th century
A.D. A glass dish, probably Roman, of about the same
period is more vaguely recorded from Honan! and other
eatly glass from China may include Western elements,
though much more archaeological and spectrographic
analysis is required before the extent of these elements can
be estimated. If we accept the evidence of the Pei-shih,
a relatively late version of earlier material, glass was not
actually made in China until the second quarter of the sth
century A.D., when the process is said to have been introduced
by traders from the country of Ta-yuch-chih, bordering on
the north-west of India. Another version has it that the
traders came from India itself. On the other hand it has
been urged that certain Chinese glass beads resembling
Western ‘eye-beads’, with the glass body inlaid with white
rings and a coloured centre, indicate a partial introduction of
the technique at a somewhat earlier date. Glass fragments
found by Sir Aurel Stein in central Asia may indicate the
line of approach, but once more the absence of precise
detail is baffling. When China is once more accessible the
whole problem is deserving of further attention. Mean-
while Rome in China remains the province of history—
particularly of Chinese history—rather than of archaeology.
1 For these and other Western objects from China, the most accessible source
is C. G. Seligman, ‘The Roman Orient and the Far East’, in Antiquity, XI
(1937), sf.
 XV ~REDROSEEGE

Re trade, diplomacy and ‘drift’ have now been

traced, in however sketchy a fashion, far beyond the
Imperial frontiers in Europe, Africa and Asia. No
attempt has been made to re-traverse in detail the ground
covered in the recent past by Charlesworth, Warmington
or Eggers; on their basic work, and on their ancient author-
ities, the present book is founded. But here and there it has
been possible to add to their material, to give it perhaps a new
actuality and a new perspective. The French discoveries in
Afghanistan and Indo-China, those of the British at Taxila and
Arikamedu or of the Italians in the Fezzan, have necessitated
a partial rewriting of the story. Only in Europe itself,
to-day harassed and sub-divided, has discovery almost ceased
for the time being.
In conclusion, let us glance at the evidence as a whole and
estimate very briefly the value of the new material. Roman
trafficking with lands outside the Empire was founded
primarily on the supply of five commodities which were
woven into the fabric of Imperial culture, and were essential
in one way or another to the Imperial way of life. Free
Germany produced the amber which was already an in-
tegral part of the equipment of southern Europe before the
Empire was born. -From tropical Africa came the ivory of
infinite domestic use, either across the Sahara or by way of
the Red Sea. Southern Arabia yielded the frankincense
which had long been sought by the Pharaohs and the
Achaemenids and, with gold and myrrh, was offered at
Bethlehem. In peninsular India grows the abundant pepper
which has for more than 2,000 years mitigated the cooking
of the Western World. And China had a monopoly of silk
until the 6th century a.p. Amber, ivory, incense, pepper,
silk were the mainsprings of Roman long-range trade. The
principal routes and markets were determined by them; a
 RETROSPECT 177
map of them is substantially a map of Imperial commerce as
a whole. “Into these routes and markets were fed, as was
inevitable, a miscellany of other goods—pearls, semi-precious
stones, tortoiseshell, muslin, skins, spices—which varied
from place to place and served to enlarge business and
make up cargoes. But all these things were supplementary
to the Great Five, and an appreciation of this fact at once
simplifies and rationalizes the complex story which other-
wise is liable to baffle the reader.
Then, again, it may sometimes be worth while to dis-
tinguish, where possible, between deliberate, organized
long-range trade and what may best be called “commercial
drift’. Trade in the five staple commodities was certainly
organized. The commercia of the Baltic amber coast, the
‘treaty ports’ of the Orient, are illustrations, but the matter
is not in doubt. How far, on the diplomatic level, the
machinery of trade was an affair of the Imperial government,
and how far it rested with the individual prestige and
experience of the Roman shipping-companies is less certain.
In the home waters of the Red Sea, direct government
control was exercised; the centurion of Leuké Komé (p. 116)
is good proof, and the licensed tax-collector Annius Plocamus
exercised rights far round the Arabian coast (p. 128). It
may be that the “Temple of Augustus’ at Muziris (p. 121)
indicates an official arm long enough to reach the Malabar
Coast. But it may also be suspected that much of the more
distant negotiation rested with commercial enterprise on a
less formal basis, as in the earlier days of the East India
Company. The so-called ‘embassy’ to the Chinese king in
A.D. 166 (p. 174) has an unofficial ring about it, in spite of its
pretension. And, as we have seen, the presence of sporadic
Western goods at Oc-eo in Indo-China suggests mere drift
in an intercourse essentially Indian rather than Roman in its
structure. To drift likewise may be attributed the pots and
coins which straggled on to the high plateau of the central
Deccan and were sometimes copied there by local craftsmen.
Most of the Roman goods from Free Germany were doubtless
178 ROME BEYOND THE IMPERIAL FRONTIERS

the product of drift from market to market or individual

to individual, even when all allowance is made for diplom-
atic gift or military and tribal movement.
Each major discovery during the past 20 years has con-
tributed, not only to the material content of our knowledge,
but inferentially to the significant enlargement of our
historical perspective in respect of these trans-frontier
contacts. In the Fezzan the Germa monument, for all its
loneliness or even because of it, has about it the suggestion of
something more than casual trading, and has here been
recognized tentatively as a relic of a political agency dis-
guised (as so often nowadays) in the shape of a technical
mission (p. 106). Arikamedu or Podouké, on the Bay of
Bengal, has shown how, in the time of Augustus and
Tiberius, Roman wine and table-wares were already reaching
the far-off Coromandel coast in quantity, and the map of
early Roman coinage across south India has pointed to the
trans-peninsular route whereby they came from the harbours
of Muziris or Nelcynda. The Arikamedu excavations
have revealed for us how at this time a new commercial and
industrial town sprang up here on the site of an insignificant
village, and, with the help of the Periplus and Ptolemy, have
illustrated the astonishing fashion in which under the
Principate, the shores of India—eastern no less than western
—lit up with the beacons of a new international trade. And
the Tamil literature has shown us vividly something of the
foreign quarters of these harbour-towns, with their ‘abodes
of the prosperous Yavanas’ and their ‘sailors from many
lands’. The Yavana sailors must mostly have been Egyptian
Greeks and Arabs, but we may suppose that here and there a
more or less resident freedman represented the interests of a
master sitting afar off in one of the shipping offices of
Alexandria or Ostia, Myos Hormos or Berenice. To such
accredited freedmen doubtless devolved the patient task of
negotiating with local rajas and producers.
Two thousand miles further north, the excavations of Taxila
and Begram have produced a different but no less revealing
 RETROSPECT 179
picture: of a rich trade in transit to and from further Asia,
shedding samples en route in market or douane, but not
chaffering to any great extent in these lands of passage. The
two famous sites are complementary, and together con-
stitute a formidable addition to the older evidence. Then,
finally, the excavations of British, French, Japanese and
other archaeologists in Indo-China and central Asia have
hinted at possibilities rather than produced finite information.
There further evidence must be awaited.
From all this trafficking emerges a question of values which
deserves some attention in the long view of history. To
what extent were Western ideas diffused in the course of this
vast diffusion of Western commodities? The importance of
trans-frontier trade as a reflection of Roman enterprise and
determination is manifest. Had it any more enduring
consequence as a social or cultural stimulus in the lands to
which it penetrated? More often than not the answer to this
question must be ‘No’. The Free Germans learned in one
way or another to use Roman weapons and equally ac-
customed the Romans to the use of German or Scythian
arms. The glamour of the Roman world, brought to their
doors by trade and war, drew the migrating tribes and in
some measure directed the refashioning of Europe; but it
cannot be said that in its homes Free Germany was greatly
influenced by contacts with Roman culture. The restless
Garamantes of Africa may have been anchored temporarily
in the interests of Rome by an imposed agricultural system
and occasional trade, but their romanization, such as it
was, had no long-term effect upon the cultures of inner
Africa. The far more extensive contacts with South India
have been a blessing to the archaeologist in search of fixed
points for the dating of associated Indian cultures, but had
no appreciable influence upon those cultures themselves.
Only in the north, in north-western India (West Pakistan)
and Afghanistan, have we found evidences of a fructual
contact between West and East. There the combined
accidents of the appropriate aesthetic and religious needs of
180 ROME BEYOND THE IMPERIAL FRONTIERS

the contemporary Buddhism, of a wealthy and powerful

native régime favourable to that Buddhism, and of the
diversion of much of the trans-Asian traffic of the West
through the region by Parthian intransigence, brought the
arts of the Mediterranean and of India together in a rare
mutual understanding. Unless it be in the mingling of
Muslim and Hindu traits in the medieval architecture of the
same sub-continent, it is difficult to find another occasion on
which so close an integration of two essentially alien tradi-
tions has been achieved. Disputation as to the precise
manner of that integration—and disputation will continue
until more exact evidence is forthcoming—cannot obscure
the magnitude of the fact. And it is tempting to insist rather
upon this positive instance of the transmission of ideas than
upon the more negative transmission of pots and pence with
which the bulk of our study, as of such studies generally, has
been concerned.
Outside the immediate range of commercial and cultural
values, a brave attempt has been made by an American
historian? to fasten upon this international trade the responsi-
bility for the major wars and migrations in Europe and
Asia during the 1st centuries B.c. and a.p. and, by inference,
over an indefinitely larger period. Within a selected time-
bracket (58 B.c. to A.D. 107), the writer in question has
tabulated, region by region from Britain to Cambodia, the
known wars and disturbances, and has observed that every
barbarian uprising in Europe followed the outbreak of war
either on the eastern frontiers of the Roman Empire or in
the “Western Regions’ of the Chinese. What is the signi-
ficance of this coincidence? It is concluded that the corres-
pondence of wars in the East and the invasions in the West
as due to interruptions of trade, with Chinese aggression as a
primary agency and Roman ageression as a secondary one.
‘When war occurred on the routes in the Tarim basin,
disturbances broke out in Parthia and either in Armenia or
on the border of Syria. Evidently, then, war in the Tarim
1F. J. Teggart, Rome and China (Univ. of California, 1939).
 RETROSPECT 181

occasioned an interruption of traffic on the silk route, and this

interruption aroused hostilities at points along the route as
far west as the Euphrates. It seems highly probable, for
example, that the invasions of Armenia by the Parthians,
while Armenia was controlled by Rome, were inspired by
the suspicion that the Romans had succeeded in diverting
the movement of commodities from Central Asia to some
route which avoided Parthian territory. But these secondary
or derivative wars . . . brought about new interruptions of
trade, and thus led to new wars in more and more distant
areas. So interruptions of traffic on the Black Sea stirred up
peoples north of the [lower] Danube, and the long train of
disturbances ended finally in collisions of the barbarians with
the Roman legions on the Rhine. Consequently, it is to
be seen that people in no way concerned with the Silk
Route might yet be connected with the interruptions of trade
on that route through the hostilities which the interruptions
precipitated between Parthia and Rome....’ In this theory
the Silk Route and its extensions become an instrument of
High Tragedy, of a forthright kind which by its very
simplicity is suspect as an explanation of the complex work-
ings of history. Other simple solutions of folk-disturbance
have been sought from time to time in factors such as
climatic fluctuation, evolving economy, overcrowding,
personal or national ambition. These are facets of the
problem, variously emphatic in varying circumstance; they
are not its core, which is likely enough to lie embedded
beyond the superficial view of history. Amongst them,
trade-routes and commodities may be sufficiently stressed
without the implication that history must be regarded
persistently through a shop window.
Te.
 SEEECT BIBLIOGRAPHY
GENERAL
M. P. Charlesworth, Trade-routes and Commerce of the Roman
Empire (Cambridge, 1926).

EUROPE

O. Brogan, “Trade between the Roman Empire and the Free

Germans’, in the Journal of Roman Studies XXVI (London,
1936), 19sff.
H. C. Broholm, Danmark og Romerriget (Copenhagen, 1952).
H. J. Eggers, Der rémische Import im freien Germanien (Atlas der
Urgeschichte 1, Hamburg, 1951).
H. Shetelig and H. Falk, Scandinavian Archaeology (Oxford, 1937).

AFRICA

G. Caputo and others in Monumenti Antichi XLI (Rome, 1951).

M. Reygasse, Monuments funéraires préislamique de l'Afrique du
nord (Paris, 1950).
ASIA

M. P. Charlesworth, ‘Roman Trade with India: a Resurvey’, in

Studies in Roman Economic and Social History in Honour of
Allan Chester Johnson, ed. by P. R. Coleman-Norton
(Princeton Univ. Press, 1951).
W. W. Tarn, The Greeks in Bactria and India (Cambridge, 1951).
E. H. Warmington, The Commerce between the Roman Empire and
India (Cambridge, 1928).
R. E. M. Wheeler, ‘Roman Contact with India, Pakistan and
Afghanistan’, in Aspects of Archaeology, Essays presented
to O. G. S. Crawford, ed. by W. F. Grimes (London, 1951).
 INDEX
AACHEN, ‘Hemmoor’ buckets from, Arne. les 27:
79 Arretine potters and ware, 37, 53,
Abalessa, tomb of Tin Hinan at, 107 129, 148, 149
Abyssinia, trade from, 116 Asciburgium, 16
Achilles, on cup from Hoby, 37 Asia Minor, 37, 69
Aden, Gulf of, 116, 117, 127 Asoka, 131
Adulis, 115, 118 Attila the Hun, 93
Aethiopians, 104 Aurea Chersonesus, 123
Afghanistan, 154-71 Aurigemma, S., 106
Agaritic muslins, 148
Agisymba, 105
Ahenobarbus and the Hermunduri, 17 BaB EL-MANDEB, 116
Aix-en-Provence, sculptures from, Bactra, 156,
136 Bactria, 120, 124, 172
Akota, Baroda State, jug from,152 Bactrians, 135
Alamanni, the, 49, 53, 54, 61, 62, 79 Balkh (Bactra), 156
Alaric the Goth, 122 Ballomar, king of the Marcomanni,
Alexander, king of Epirus, 131 21
Alexandria, 37, 69, 116, 155, 164, 170, Baradez, J., 95
171 Barbaricum, Sind, 118
Algeria, 95 Bardo, Musée du, Algiers, 109
Almgren, O., 20, 91 Baroda State, bronze jug and cameo
Amaravati, rouletted pottery and from, 152
Roman coins from, 151 Barygaza (Broach), 117, 120, 124,
Amator, potter, 89 T2051 3Ae TIO nS Oy 172
Amber, 3, 14, 23, 34, 42, 55, 92, 176 Beads, 175
Amphorae, 100, 102, 103, 148, 149 Begram hoard, Afghanistan, 158,
Andhra Empire, 140 162ff., 178
Anglesey, Iron Age hoard from, 56 Beni Ulid, 107
Ankara, inscription of Augustus at, 15 Benoit, F., 136
Annius Plocamus, 128, 177 Berenice, 112, I15
Antigonos Gonatas, king of Mace- Bjorska, Vastmanland Province of
donia, 131 Sweden, 53
Antiochos Theos, 131 Black Sea, 12, 18, 58, 67, 71, 72, 82,
Antoninus Pius, Indian embassy to, 86, 93, 155
135 Boats, 58, 60
Apollo Grannus, bucket dedicated Bober, P. F., 135
to, $3 Bohemia, 15, 18, 19, 20, 21, 35, 67,
Apologos (Obollah), 117 75, 85, 91, 105
Aquileia, 14, 21, 37, 73 Boii, the, 19
Arikamedu, S. India, trading-station Bolin, $2323,.63;,0%
atI20,s0ATentASs, s100.078 Bombay, 124
Arminius, 20, 54 Bonn, 16
186 INDEX

Bornholm, 29, 86, 93 Chariomerus, king of the Cherusci, 10

Borstel, Saxony, terra sigillata from, 89 Charix Spasini, 117
Bosco Reale, hoard from, 37, 40, 44 Charlesworth, M. P., 135, 176, 182
Bows, 58 Charsada, Peshawar Plain, 161
Brahmagiri, Mysore State, 151 Charydes, the, 16
Brandenburg, 21, 68 Chatti, the, ro
Broach (Barygaza), 117, 120, 124, Cherusci, the, 10
PDOs As 13On lhs Ovni) Chieftains’ burials, 30, 31, 32, 44,
Brochs, Scottish, 96 70, 84
Brondsted, J., 48 @hinay 359137401 50n 157905en 170
Brogan, Mrs. Olwen, vii, 182 ‘Chouchet’ tombs, 111
Broholm, H. C., vii, 182 Chrysé the Golden, 123, 172
Bronze vessels, 28, 30, 31, 72-84, Cicero, I
86; 1550 52, 103 Cilicia, 26
Brooches, 21, 38, 40, 46, 52, 91 Cimbri, the, 16, 29, 56, 61, 91
Buddhist art, 165 Cintugnatus, potter, 89
Buddhist shrines and monasteries, 160 Cintusmus, potter, 89
Bullae from India, 152 Ciosmy, near Lodz, terra sigillata
Bu Ngem, 107 from, 89
Burgundians, 24, 29, 68, 80 Claudius Gothicus, 26
Byrsted, Jutland, cups from, 70 Cn. Trebellius Romanus, 37
Coffins, 36, 39, 41, 46
Coimbatore, coin-hoards from, 143,
CAESAR, JULIUS, 7, 18, 21, 56, 61 145
Calliena (Kalyana), W. India, 121, Sorin, Gh, Ih, Pig, Bp, Dele Av, IY air.
124, 125 AOWASs 95250595 103-551 025 lOve Elite
Camara, S. India, 123 120, 122, 130, 137) LAl, 152
Cameo from Karvan, Baroda State, Colchester, Arretine pottery from, 89
©52 Colchi (Kolchoi), pearl-fisheries at,
Campania, 14, 37, 44, 75 123, 147
Cape Comorin, 123, 144 Cologne, 79, 85
Cape Delgado, 112 Comitialis, potter, 72, 89
Cape of Spices (Cape Guardafui), 112 Coptos, 116, 128
Capua, 43, 73, 84 Cornelius Balbus, campaign of,
Caputo, G., vil, 97 against the Garamantes, 103
Carbuncles, 106 Coromandel Coast, 131, 145
Carnelian of Rajpipla, 154 Cranganore (Muziris), 121, 123, 124,
Carnuntum, 9, 14, 23, 72, 75, 84, 89 132, 143, 177
Carpathians, 18 Crimea, 26
Casal, J. M., 146, 147 Cunobelin, 2, 53
Caspian Sea, 155 Cwrles| 2
Catualda, 20 Cydamus (Ghadamas), 107
Cernunnos, 135
Ceylon, 123; 124; 128, 134,137, 141, Dacia, 10, 12, 18
144, 145, 147 Dambitzen, Kr. Elbing, West Prussia,
Chandravalli, Mysore State, 140, 151 terra sigillata from, 89
Chaoucre, Dep. Aisne, silver bowl Damiirica (S. India), 117, 121, 124, 126
from, 7I Danzig, 15; 235.245) 26,735 84085
 INDEX 187
Dar-es-Salaam, 112 Frisian coast, 49, 50, 90
Decebalus, king of Dacia, 10 Frisians, the, 16, 92
Demetrias, Sind, 132 Frobisher, 122
Dikshitar, V. R. Ramachandra, 132 Fulda, river, 90
D195 850,10; 17, 22,135 Fyen, 27, 32, 59, 68, 70, 81
Dionysus as emblema of silver dish
from Taxila, 159
Dioscoride (Socotra), 117 GALICIA, 18
Dnieper, river, 18 Gallehus, Sleswig, drinking-horns
Dniester, river, 18, 83, 84, 86 from, 72
Dollerup, Jutland, “Chieftain’s burial’ Gandhara art, 166, 168, 173
at, 38, 40, 70, 71 Ganges, river, 123, 124, 144, 147,
Drenthe, Roman statuettes from, 52 172
Dresden, 15 Garama (Germa), 103
Drinking-horns, 21, 30, 37, 38, 39, Garamantes, the, 103, 104, 105, 106,
42, 43, 72, 86 179
Drusus, brother of Tiberius, 52, 90, 92 Gautier, E. F., 109
Dzwinogrod, Krets Bobrka, Poland, Gebel es-Soda, 107
terra sigillata from, 89 Gems from India, 146, 160
Germa (Garama), 98ff., 178
Germanicus, 52, 54
East AFRICA, I12-14 Ghadames (Cydamus), 107
‘Eastland’ type of bucket, 76 Gheria, 107
Eggers, H. J., vil, 18, 31, 40,42, 63, Ghirshman, R., 162
75, 78, 80, 82, 83, 86, 91, 176, 182 Giffen, Professor A. E. van, 49, 50
Elbe, river, 17, 19, 22, 35, 78, 90 Gile, near Oslo, bronze bowl from,
Elenius, potter, 89 82
Embassies to China, 174, 177; from Gispersleben, Thuringia, terra sigillata
India, 133, 134, 135 from, 90
Emporia, 124 Glass, 4, 42, 71, 84-7, 99, I10, 118,
Ems, river, 54 120,122, 147,450, 158, 163, £75
Engelhardt, C., 56, 58, 59 Gold, 64ff., 93, 141
Entremont, Provence, sculpture Gold vessels, 72
from, 136 Gondopharnes, king of Taxila, 133
Erfurt, 22 Goodchild, R. G., 95
Erythraean Sea, 112ff., 115ff. Goths, the, 10, 12, 24, 26, 27, 29, 58,
Etesian winds, 126 72, 93
Eudaemon Arabia (Aden), 116, 117 Gotland, 21, 26, 27, 29, 87, 90, 93
Euphrates, river, 117 Gréningen, Roman statuettes from,
Euripides, 37 §2
Eye-beads, 175 Guimet, Musée, 161, 165
Ezinge, Holland, 50

HAAGERUP, FINLAND, SILVER CUPS

FEZZAN, THE, 97-107 FROM, 70
Frankincense, 176 Hackin, J., 162
Franks, the, 53, 62 Halle, 15, 45
Friesland, 52 Hameln, 16, 90
 188 INDEX
Hammada, the, 98, 105 Julianus, 9, 14
Hanover, 36, 44, 46, 54, 68, 88, 89 Julius Maternus, mission to Agisymba,
Harpocrates, statuettes of, 158, 164 104
Helzendorf, Hanover, cemetery at, 46 Junius Alexius, lamp-maker, roo
Helmet, 57, 59 Junnar, W. India, 132
‘Hemmoor’ type of bronze bucket, Jutland, 7, 14, 15, 16, 22, 27, 28, 29,
47, 53» 545 78, 82 39, 41, 52, 70, 71, 80, 86, 87
Herculaneum, 14
Hercules, 69, 164
Hermunduri, the, 9, 17, 19, 22, 23 KaBuL Museum, 165
Hildesheim hoard, Hanover, 36, 44, Kalyana, W. India, 121
54 Kanakasabhai, V., 132
Hill, Sir George, 142 Kanishka, 162, 169
Hippalus, 126ff. Kapisa, Afghanistan, 162, 165
Hirth, F., 154, 174 Karli, W. India, 132
Hiuan-tsang, 162, 163 Karvan, Baroda State, cameo from,
Hjortspring, Danish Slesvig, peat-bog 152
deposit from, 56, 59 Khyber Pass, 156
Hoby, Laaland, ‘chieftain’s burial’ Knapke, W., 54
at, 36, 44, 70 Kolchoi (Colchi), pearl-fisheries at,
Hoggar, the, 97, 108 123, 147
Hgjsted (Rumperup), Denmark, Kolhapur, bronzes from, 151
burial at, 40 Kondapur, bullae and pottery from,
Holtgaste, Hanover, terra sigillata 151s 52
from, 88 K6nigsberg, burials near, 14
Holubic, near Prague, silver cups Korea, glass from, 4, 175
from, 71 Kushana empire, 142, 156
Honan, glass from, 4, 157 Kvalsund, Norway, peat-bog de-
Hoxter, 16 posit from, 60
Huns, the, 10, 12, 22, 27, 94
Hyderabad, Sind, 118
LAALAND, 27, 36, 41, 70
Lahn, river, 16
Iliad, the, 37
Laisack, burial at, 47
Illyricum, 19
Lamps, I00, I10, 147, 150, 158
Incense, 117, 176
L. Ansius Diodorus, 43, 84
Indo-China, 172
Laocoén on Gandhara sculpture, 161
Indragnitta of Demetrias, 132 Lapis lazuli, 47, 120, 154
Indus (Sinthus), river, 118, 126, 127
Lausitz, burials at, 80
Intaglios from India, 146, 160
Leg Piekarski, Poland, silver cups
Ireland, 2
from, 71
TvOry, 07, 105,71135 116. 122, 145, Leipzig, 15, 22, 46
163, 176
Lepcis Magna, 97, 104, 107
Iyengar, P. T. Srinivas, 132
Leuké Komé, 116, 125, 177
Leuna, Saxony, silver cups from, 7I
JOHANSEN, K. Frus, 36 Libya, 96
Juellinge, |Denmark, ‘chieftain’s Ligurian sculpture, 136
burial’ from, 41, 42 Linz, 15
 INDEX 189
Lippe, river, 16, 90 Mizda, Tripolitania, 107
Long-houses in Denmark, 48 Moldau, river, 15, 19
Loot, 53 Mollerup, Jutland, silver cups from,
Long-Xuyen, Indo-China, 173 71
Liibsow, ‘chieftains’ burials’, 32, 34, Mombasa, 113
35, 43, 70, 71, 75, 84 Mons Ater, Libya, 107
Moscha, 117, 125
Muller, Sophus, 41
McCrinDLg, J. W., 172 Murzuch in the Fezzan, 98
Madura, S. India, 133 Muzavern2, 103, 160s 117,
Maés (Titianus), 155 Muziris (Cranganore), 121, 123, 124,
Magas, king of Cyrene, 131 132, 143, 177
Magdeburg, 22 Myos Hormos, 115, 125, 129
Mail, shirts of, 56, 59
Main, river, 16, 19, 90
Mainz, 19, 22
NAISSUS, BATTLE OF, 26
Maiuri, A., 135
Makran, 127 Naristi, the, 19
Malabar coast, 137, 143 Neckar, river, 19
Malaya, 172 Nelcynda, S. India, 121, 123, 143, 144
Malleret, L., 173 Neptune, statuette of, from India, 151
March, river, 14, 15, 35, 73 Nicolaus of Damascus, 133
Niemen, river, 23
Marcomanni, the, 8, 9, 10, 14, I5, 17;
Nikon, E. Africa, 112, 114
18, 19, 20, 22, 35, 67, 73, 75> 70, Nordrup, Zealand, silver cups from,
85, OI, 105
Marcus Aurelius, 21, 67, 68, 92, 93, 71
‘Nordrup-Jesendorf’ glass, 85
174. INoricum, 14, 20) 21, 22, 41
Markanan, S. India, 123
Norling-Christensen, H., 29
Marmaridae, the, 103
Nubia, 116
Maroboduus, king of the Marco-
Nydam, peat-bog deposit at, 58, 60
manni, 8, 19, 20, 21, 73, 75, 91, 105
Marshall, J., 143, 157
Masalia, India, 123
Maski, India, rouletted pottery from, OBOLLAH, I17
151 Ocelis, 116, 117
Massilia (Marseilles), 35 Oc-eo, Indo-China, 172, 177
Mathew, Rev. Gervase, 114 Oder, river, 18, 32, 35, 74, 85
Mattingly, H., 113 Oea (Tripoli), 97, 104, 107
Mecklenburg, 80, 81, 85 Oland, 27, 93
Mehrum, near Diisseldorf, terra @rsnes-Christensen, M., 39
sigillata from, 89 Ommana, 118
Meissen, 15 Opone (Ras Hafun?), 112
Menander, 131, 132 Orange, Provence, 56, 61
Menouthias (Zanzibar?), 112 O Riordain, Professor Sean, 2
Meredith, David, 128 Orosius, 61
Meroe, 116 Oslo, 30, 82
Minden, 16, 90 Ostropataka, Czechoslovakia, silver
Minnagara, Sind, 118 cup from, 72
190 INDEX
PAcE, PROFESSOR, 97 Ptolemy, 17, 104, 112, 124, 131, 133,
Paderborn, 16 134, 14, 147, 172
Pakistan, 154-71 Puhar, S. India, 133
Palaesimundu (Ceylon), 123 Pyralae Islands, 112
Palmer, |... Bz 1¥5
Palmyra, 170
Pannonia, 14 QuaDI, THE, 10, 15, 18, 19, 22, 35, 92
Parthians, the, 137, 154, 158
Passarge, river, 24
Patale or Patala, 118, 127 RAETIA, 16, 17, 19, 22, 49, $4, 61, 80
Pausanias, 140 Rajpipla, carnelian from, 154
Pearlsaan 22501234150 Ras Fartak, Arabia, 127
Peat-bogs, 31, $4, 81 Ras Hafun, 112
Pepper, 3, 121, 122, 130, 132, 145, Ravenna Geographer, 121
176 Red Sea, 112, 115, 125, 127
Periplus of the Erythraean Sea, 112, Reygasse, M., vu, 109
115-25, 126, 147, 148, 154 Rhapta, 112
Pernice, E., 43, 69 Rhenish beaker, 48
Persian Gulf, 117 Rhys Davids, T. W., 131
Peshawar Plain, 161 Rilled bronze bowls, 30
Petra, 116 Rodd, F. Rennell, 105
Peutinger Table, 121 Roquepertuse, Provence, sculpture
Philoctetes of Euripides, 37 from, 136
Piracy, 122 Rostovtzeff, M., 170
Pliny the Elder, 9, 23, 37, 103, 104, Rouletted pottery, 140ff.
TiS eT O Wel 2a Ones 4h Oe wAT Ruhr, river, 16
Plocamus, 134, I4I Runic alphabet, 27
Plutarch, 134
Podouké (Podouke), 123, 124, 129,
147, 178 SAALE, RIVER, 22, 45
Pomerania, 70, 82, 86 Sabratha in the Hadramaut, 97, 117
Pompeii, 14, 43, 135 Sahara, the, 95-111
Pondicherry, S. India, 123, 129, 145, St. Thomas, 133, 143
147 Samian, see Terra sigillata.
Port Durnford, E. Africa, coins from, San, river, 86
113 Sarapion, E. Africa, 112
Porus, 133 Saxony, 21, 68, 71, 89
Poseidon, statuette of, 151 Schleswig-Holstein, 68
Posen, 21, 68 Schulz, W., 46
Postumus, emperor, 62 Scotland, 2
Pregel, river, 23 Scythians, the, 118
Priam and Achilles on cup from Seligman, C. G., 175
Hoby,37 Selling, Dagmar, 87
Priscus, 93 Semnones, the, 16
Probus, 68, 71 Septimius Flaccus, expedition of, to
Prodbaba, near Prague, terra sigillata the Aethiopians, 104
from, 90 Septimius Severus, 23, 64, 66, 95
Ptolemais, 116 Sergi, Professor, 98
 INDEX I9I

Seyrig, H., 170 Terps, 49, 50

Shetelig, H., 27, 72 Terra sigillata, 17, 47, 52, 72, 87-90,
Shields, 57, 60 100, 103
Siam, Gulf of, 172 Teutons, the, 56, 61
Sigerus, W. India, 127 Theiss, river, 18
Silenus, silver emblema from Taxila, Thinae, 172
159 Thompson, E. A., 93
Silesia, 15 Thorsbjerg, Denmark, peat-bog de-
Silius, 37 posit from, 56, 60
Silk watn A) 920, -121;.122, 123,150; Thuringia, 9, 16, 17, 22, 67, 68, 78,
TSA eSOs 172, 170s TOT 84, 85, 89, 90
Silk Route, 154, 181 Tibesti, 97
Silver vessels, 36, 40, 43, 68-72, 85 Tifinagh inscriptions, 110
Sinthus (Indus), 118, 126, 127 Tin Abunda, 98
Sisupalgarh, India, 151 Tin Hinan, 107-11
Slaves, 97, 113, 118 Titianus (Maés), 155
Slesvig, 32, 72 Toulouse, 56
Socotra, 117 Tranquebar, 143
Sogne fiord, 74 Tripolitania, 97
Sopatma, S. India, 123 Trojan Horse, 161
Ope An Cut 70, 171 Trondheim, 4, 30, 44, 46, 80
South Russia, 12, 18, 72, 93 Trumpets, 55
Spurs, 40, 46 Tuareg, 109
Statuettes, 52, 135, ISI, 164
Stavanger, 30 UKRAINE, 84
Stelae in the Fezzan, 99, 100, 102
Stockholm, 30
Storedal, Norway, 32 VAASA, FINLAND, BURIALS AT, 46
Strabo, 7, 17, $6, 91, 100, 126, 129, Valerius Festus, expedition of, 104
130, 133, 134, 144 Valloby, Zealand, silver cups from,
Stucco, 160f., 164, 168ff. 71, 82, 89
Suebi, the, 7, 18 Vandals, 24, 68
Sundsvall, Sweden, 80 Varpelev, Zealand, glass and silver
cup from, 71
Varus, 3, 20, 54, 67
Tacitus, 7, 8, 9, 17, 19, 26, 34, 38, Vetera (Xanten), 16
41, 45, 48, 64, 66, 92, 140 Vimose, Denmark, peat-bog deposit
Tafilalet, 108 from, 59
Takamat, 109 Vippachedalhausen, Kr. Weimar,
Taprobane (Ceylon), 123 Thuringia, terra sigillata from, 89
sara) Vistula, river, 24, 68, 82, 83, 86
Tarn, Sir W. W., 128, 131 Wess, ©).4 30, Fit
Tarvagus, potter, 89
Tashkurgan, 3, 155, 157
Taxila, Punjab, 118, 233,137, 157, Want EL-AGIAL, 98, 105
168, 169, 178 Ward Perkins, J. B., 95
Teggart, F. J., 180 Warmington, Professor E. H., 127,
Tensfelderau, coin-hoard from, 67 128, 134, 143, 176, 182
192 INDEX

Waywood, Captain C. W., 113 Wine, 33, 45, 118, 120, 133, 147, 149,
Welwyn, 2, 38 158
Werner, J., 38 Wittislingen, Bavaria, huts at, 49
Wieser, river, 16, $4, 74, 90 Woollen clothing, 58
‘Westland’ type of bronze bucket, 86
YAVANAS, 132, 133, 178
Westphalia, 16
Yonakas, 132
Wetterau, river, 22, 90
White Huns, 163, 166, 169 ZANZIBAR, I12
White Russia, 18 Zealand, 27, 29, 71, 78, 82, 84, 85,
Wilkens, H., 52 86, 87, 89, 92
 authoritative
com prehensive
elegant

& Collegiate\
%

of the Englisha
Zia

The New Collegiate is a superb modern diction-

ary of very convenient size. It is based on the
creat We bster’s New International to providea
handy -size dictionary, of the same high standard
of scholarship, for students and the general
reader. The definitions are those of the New
International modified to obtain greater brevity
and directness. Wherever they are needed
phrases and sentences have been given to illus-
trate the meanings.
The entries concerned with the vocabularies
of specialized subjects, from Accountancy to
Zoology, benefit from the remarkable accuracy
and completeness of the New International in
the preparation of which over 200 specialist
authorities have been employed as editors.
Over 125,000 entries - 2,300 illustrations + 1,230
pages + Buckram Binding + Pronunciation of every
entry + Synonyms . Full etymologies . Bio-
graphical Dictionary + World Gazetteer

This cdition may not be supplied to the U.S.A. and its

Dependencies, Canada, Newfoundland, Australasia.
 .

some BELL books

ROMAN GAUL
by OLWEN BROGAN, F:s.a.
- TIMES EDUCATIONAL SUPPLEMENT: ‘This book fills a curiously unexpected want.. .
There is no French work with anything like the profusion of excellently chosen draw-
ings, and photographs . . . Sections show an admirably wide and personal knowledge of
the remains . . . Arles, Frejus, Nimes, and Lyons, could hardly have been better
handled. This reviewer has been delighted by the easy graceful style in which it is
written.” 215. net

LOGAL RECORDS
Their Nature and Care
edited by LILIAN’ J. REDSTONE, .B.£., B.a.
and F. W. STEER, F:s.A., F.R.HIST.S.
THE C.C.A. GAZETTE: ““The method of presentation, the scope of the book as a’ hand-
book of archive practice, and the description of the work on local records which is now
going on up and down the country should result in its being read not only by local
government members and officers but by all who have in their charge the care of
family papers and archives.” 255. net

THE HERITAGE OF EARLY BRITAIN

by the lae MARTIN CHARLESWORTH and other historians and
~ archaeologists of the University of Cambridge
TIMES LITERARY SUPPLEMENT: ““This is a delightful little book. It attempts so little—
to stimulate a wider interest in the prehistory and early history of Britain—and it
achieves so much.” Well illustrated. 125. net

* IN PREPARATION *

ANCIENT LANDSCAPES
IN EUROPE AND ASIA
Studies in Archaeology and Air Photography
by JOHN BRADFORD, .a,, F.s.a.
University Demonstrator and Lecturer in Ethnology in the University of Oxford
This book describes new discoveries made by the author from air photographs, and
their importance for the study of archaeology and history. it is illustrated by about
sixty plates of magnificent air photographs, selected from many thousands, and the
great majority are here published for the first time. Price about 30s. net

MYTH OR LEGEND?
The Flood, Atlantis, the Trojan War, Tristan and Isolt—is each of these famous stories
mere imagination—a myth—or are they based on facts, facts distorted or exaggerated
into a legend? This book is based on twelve lively talks by experts, broadcast on the
Home Service. Jlustrated. Probable price ros. 6d.

LONDON. -G BLE AND sO hs Seer