CHAPTER 3

Local Networks and Long-distance Trade:

The Role of the Exchanges between
Sri Lanka and India during the
Mediterranean Trade
A R I A N E de S A X C É

Recent archaeological fieldwork and studies in Sri Lanka and South

India have opened the way to a re-evaluation of the role of local
networks of circulation and trade between South Asia and the
Mediterranean. Instead of considering that the Egyptian Greeks or
Romans handled the trade by themselves, inaugurating routes and
contacts, it appears that they rather followed and re-used networks
that were existing locally long before the outburst of exchanges
between the Roman Empire and India in the early centuries of the
Common Era.
As archaeological materials show, Sri Lanka seems to have been
trading with western countries through the intermediary of Indian
merchantmen before the Common Era and during the first three
centuries CE. Then comes a shift in the trade distribution when
contacts are redirected towards Sri Lanka in the fourth and fifth
centuries CE. We would like to trace the evolution of this pattern
from the fifth century BCE to the beginning of the seventh century
CE , mainly through the archaeological artefacts.
Thus, we will first consider what kind of relations connected Sri
Lanka to India in the pre-Periplus period, including contacts with
North India and Afghanistan, as well as South India. Particular
attention will be given to the question of the circumnavigation of
the island. Secondly, we will focus on the information suggested by
54 Ariane de Saxcé
the Periplus itself, examined through the prism of network analysis,
before lastly pointing out the change of situation by the end of the
Roman Empire and after.

THE PRE-PERIPLUS TRADE BETWEEN SRI LANKA

AND THE COASTS OF INDIA

Evidence from excavations in Sri Lanka during the last two de-
cades and recently have shown ancient trade contacts with both
coasts of North India, as well as close interaction with South India.
These discoveries indicated a well-established system of maritime
exchanges along the coasts and hinterland via the estuaries of
rivers, at least from the fifth century BCE onwards. Moreover, not
only the North and Dry Zone were strongly connected to South
India but also the South and Wet Zone, with the use of circum-
navigation, which paved the way for the Romans a few centuries
later.

NORTHBOUND TRADE
Contacts with the north-west coast of India were already estab-
lished at least from the fourth century BCE, as they are mentioned
by the Greek authors accompanying Alexander,1 and diverse pieces
of evidence show them as regular after the introduction of Bud-
dhism on the island in the middle of the third century BCE. Apart
from chronicles and literary accounts, archaeological material, in-
scriptions and coins indicate trade relationships with Afghanistan
and the north-west coast India. Thus, the site of Anuradhapura
(citadel and Salgaha Watta 2) has yielded one sherd of Northern
Black Polished ware from period 17, dated from 360 to 190 cal.
2
BCE , that suggests a link with North India where this ceramic
type originates. In Tissamaharama,3 one body sherd and one rim

1
Starting from Onesicritus; for further textual references, see Weerakkody,
1997.
2
Coningham et al., 2006, p. 165.
3
Schenk, 2001, p. 128.
 Local Networks and Long-Distance Trade 55
sherd derived from phase c1, that is second century BCE, and 37
samples of Plain Grey ware, from phase a (fourth-third centuries
BCE ), imported as well from the north-west of India, may even
indicate contacts before the mission of Mahinda in 250 BCE.
This is in keeping with the 15 punch-marked coins, among
which two from the same period I, put to light in this site.4 Such
type of coins has been recovered from the Citadel, the Jetavana
and Abhayagiri viharas as well as from the south of the island in
Tissamaharama, and in many places in hoards. Their circulation is
testified by an early Br"ahm$û inscription of the third century BCE
referring to trade in kahapanas.5 They belong to a class which can
be attributed to the middle and late Mauryan period, so that it is
probable that they have reached the island with the first Buddhist
contacts under King A« s oka.6 Besides, the absence of any Sri Lankan
specific symbol indicates that these coins were imports from
India.7 Similarly, the Mah"a va^m« sa refers to eight kinds of pearls
presented to King A« s oka by king Devanampiya Tissa (reigning
from 250 to 210 BCE).8
The presence of imported material in a hinterland site such as
Anuradhapura in an early period was made possible by the typical
trade pattern of South India and Sri Lanka, linking inner political
centres to harbours on the coast. In that case, the site of Mantai,
on the Palk Strait, was related to the capital city of Anuradhapura
through the Aruvi Ari River, such as the southern capital of Ruhuna,
Tissamaharama, was connected to the port of Kirinda thanks to
the Kirindi Oya, or Ridiyagama to Godavaya with the Walawe
Ganga.
Among these sites, Mantai,9 Anuradhapura and Ridiyagama10
have yielded lapis lazuli beads or raw material, known for coming

4
Bopearachchi, 2006, p. 7.
5
Bopearachchi, 2002, p. 105.
6
Bopearachchi, 2006, p. 16.
7
Ibid., referring to Codrington.
8
Rajan, 2002, p. 85.
9
Carswell, 1991, p. 200.
10
Bopearachchi, 2002, p. 107; 1999, p. 14.
56 Ariane de Saxcé
from the famous source of Badakshan11 in the north-east of Afghani-
stan or from Baluchistan.12 Five beads have been discovered in
Anuradhapura SW2 from periods I, G and F (360 BCE to 600 CE):
as the stone does not occur naturally in Sri Lanka, ‘it indicates
that it is an imported item and that there must have been long-
distance trade as early as 360 cal. BCE’.13 It is also referred to among
the 300,000 beads of the Jetanava treasure and has been preserved
in traces underneath the line of geese of the ayaka of the stupa,14
attributed to the third century CE, showing a continuity of its use.
In the reverse direction, Sri Lanka had to offer its own precious
and semi-precious stones, such as those inserted in two rings dis-
covered in the city of Aï Khanoum. This Greek town of Afghani-
stan was destroyed in the middle of the second century BCE, which
gives a terminus ante quem for those jewels decorated with one
sapphire and one ruby coming from Sri Lanka, complemented
with diamonds from India.
These links with Central Asia are correlated by several inscript-
ions referring to ‘Kaboja’ or ‘Kabojikas’, which Paranavitana inter-
prets as ‘Kamboja’, a population of Arachosia in the extreme west
of the Mauryan Empire, mentioned together with Yonas, Sakas
and Gandharas in the texts. Several inscriptions from the southern
kingdom of Rohana in Hambantota disctrict refer to ‘kaboja’ or
‘kabojhiya’ as a proper name; another inscription, from Kurunegala,
close to Anuradhapura, refers to a corporation of ‘kabojikas’. Their
organization in a corporation may allude to their involvement in
trade. 15
Thus, these pieces of evidence invite us to consider that there
was an established system of exchanges between Sri Lanka, Central
Asia and the north-west of India from the third century BCE at

11
Von Rosen, 1988, p. 11: ‘the earliest literary attested source of lapis
lazuli and also the hitherto richest source of lapis lazuli is Badakshan in north-
east Arghanistan’.
12
Coningham et al., 2006, p. 377.
13
Ibid.
14
Unesco, 1993, p. 84.
15
Bopearachchi, 2002, p. 108.
 Local Networks and Long-Distance Trade 57
least. The coastal journey along the west and east coasts of India
must have facilitated stopovers and redistribution of products in
the peninsula, such as carnelian on the west coast and Rouletted
ware on the eastern side. Carnelian appears in volcanic traps of the
central Deccan plateau, where it has been a source of raw material
since Harappan times.16 When the rock disintegrates, the nodules
of chalcedony, which are harder than the matrix, are released.17 By
200 BCE, Ujjain was one production centre among many others
which are still to be identified.18 Studies on modern techniques,
that seem to be in continuity with ancient ones, suggest that the
raw material was extracted, baked and then exported to be fin-
ished on another site, as from the Narmada Valley to Cambay for
example.19 In Sri Lanka, many beads, intaglios or cameo blanks
testify to the export of this material to the coastal cities and the
inland centres. In the Jetavana treasure, among 16 intaglios, the
majority are in carnelian and, except for two of them, depict local
humped bulls or local animals. Moreover, carnelian blanks also
indicate that seal carving was practised using an imported mate-
rial.20 In Anuradhapura, a total of 102 pieces of carnelian have
been discovered, with 13 rings. They date from 510 BCE to 1100 CE,
and bead debitage is the predominant form, suggesting local manu-
facture, especially during period G (c. 200 cal. BCE-CE 130 cal.).21
Even in the south and south-west coasts, carnelian has been at-
tested, notably in Kataragama22 and Ridiyagama.23 The port of
Mantai, close to India, has also shown beads similar to those exca-
vated in India.24 Even the site of Ibbankatuwa, dated between the
seventh and the fourth centuries BCE has yielded etched carnelian

16
Roux and Pelegrin, 2000.
17
Allchin, 1979, p. 92.
18
Ibid., p. 97.
19
Ibid., p. 101.
20
Unesco, 1993, p. 83.
21
Coningham et al., 2006, p. 380.
22
Bopearachchi, 2002, p. 107.
23
Bopearachchi, 1999, p. 16.
24
Carswell, 1991, p. 202.
58 Ariane de Saxcé
beads,25 thus indicating the wide network of connections in the
subcontinent from the middle of the first millennium BCE.
The presence of horse bones and teeth in Anuradhapura26 can
be interpreted in two ways: one famous source could have been
Ferghana in Afghanistan and they would have followed the west
coast of India with the other goods mentioned above, but it can
also be assumed that they might have come through the eastern
route. Indeed, a seal discovered in Chandraketugarh depicting a
horse in a ship, and the indication of the import of horses at
Kaveripattinam in the Pa|t|tinapp"alai, suggest that ‘horses, brought
probably from the north-west of India, were shipped from Chandra-
ketugarh to various ports on the east coast of India for inland desti-
nations in South India and Sri Lanka’,27 all the more so as the east
coast is the place of expansion of the Rouletted ware, the exact
origin of which is still in debate. Ford et al.28 are inclined to believe
in one single geological source and similar techniques, whereas
Magee29 divides several groups in Ford et al. results, showing a
different origin for Rouletted ware and Grey ware in Anuradhapura.
Gogte30 and Schenk31 rather favour the hypothesis of several pro-
duction centres in the plains of North India, whether in Bengal
or in the middle Ganga plains or both.32 In all perspectives, even
though local imitations have been made, the distribution map of
Rouletted ware and its presence in Tissamaharama from the third
century BCE, indicate close cultural interactions on the eastern side
of the peninsula, even perhaps following the dispersion of the Bud-
dhist monks.33 In the south-west of the island, the site of Kelaniya,34
in a ceramic assemblage indicating an occupation from the third
25
Coningham, 2002, p. 104.
26
Ibid.
27
Gogte, 1997, p. 81.
28
Ford et al., 2005.
29
Magee, 2010.
30
Gogte, 1997, 2001.
31
Schenk, 2006.
32
Ibid., p. 140.
33
Schenk, 2001, p. 130.
34
Bopearachchi, 1999, p. 13.
 Local Networks and Long-Distance Trade 59
century BCE to the second century CE, revealed some Rouletted
sherds, for the first time in this part of Sri Lanka. Their context
was radio-carbon dated between 250 and 185 BCE for some of
them and 115 BCE to 25 CE for the rest.35 It indicates early links
between the west coast of Sri Lanka and the east coast of India.
This continuity of contacts and probable redistribution along
the coasts of India up to the north most probably took advantage
of the close links existing between Sri Lanka and the southern part
of India since the Iron Age.

INTERACTION WITH SOUTH INDIA

Culturally the similarities between Iron Age settlements in South
India and in Sri Lanka have been underlined, for example in the
early phase of Anuradhapura, where the phase J dated from 510 to
340 BCE shows strong affinities with the Iron Age cultural tradi-
tions of peninsular India,36 so that in terms of commerce these
cultural connections may have played a part in facilitating the trade
exchanges.
Cultural parallels have been shown in particular by the study of
graffiti marks achieved by K. Rajan and Osmund Bopearachchi,37
comparing the results of the excavations in Kodumanal, Tamil Nadu,
and in Ridiyagama, in the south of Sri Lanka. The graffiti under
study are thought to be clan marks, as in Kodumanal each mega-
lithic burial was associated with a special symbol engraved on all
its pottery.38 Nevertheless, in the habitational context more than
50 marks have been observed, which seems too much for clan signs
or may at least indicate some other meaning associated with the
marks. They are all post-firing graffiti, engraved most often on the
shoulder of Black and Red ware, and are found in Kodumanal
during the late phase of the Megalithic period. The similarities of
symbols—like the moon, the ladder or the swastika—between

35
Ibid., p. 20.
36
Coningham, 2002, p. 103.
37
Bopearachchi and Rajan, 2002.
38
Ibid., p. 104.
60 Ariane de Saxcé
Kodumanal and Ridiyagama, and the widespread area where simi-
lar marks are found,39 including Kelaniya and Mantai in Sri Lanka
or Korkai and Alagankulam on the Indian side of the Palk Strait,
invite us to consider close social and maybe political and economic
links between central Tamil Nadu and the southern coast of Sri
Lanka.
These economic links appear indeed in some of the archaeologi-
cal material of Tissamaharama and Ridiyagama, particularly the
beads. We have already evoked the lapis lazuli and carnelian of
Tissa; another stone deserves attention as far as movements are
concerned: the garnet. Its composition can be analysed with preci-
sion and corresponds to specific geological contexts, related either
to Sri Lanka or to India. In Sri Lanka, the production area is centred
around Ratnapura, or Ambalantota and Embilipitia, and delivers
garnets of the almandine-pyrope composition.40 The majority of
the garnets, either from Tissa or from Sri Lanka, match the com-
position of the Ratnapura raw material samples.41 But six garnet
beads of Tissamaharama/Akurugoda bear a highly almandine
composition typical of India.42 The comparison with south Indian
garnets matches with them and implies an import either of the
raw material or of the finished products.43 In Ridiyagama, the com-
mercial orientation of the site is made obvious by the high quan-
tity of punch-marked coins (and later by Roman third brass),44 for
a settlement dated between the fourth century BCE and the sev-
enth century CE,45 where more than 150 types of semi-precious
stone and glass beads have been recovered, showing many similari-
ties with south Indian typologies.

39
Ibid., p. 97: ‘such marks are reported from sites like Karur, Alagankulam,
Uraiyur, Kodumanal, Arikamedu and Korkai in Tamil Nadu and Anuradhapura,
Kantarodai, Manthai, Kelaniya, Ridiyagama and Pomparippu in Sri Lanka’.
40
Périn et al., 2007, pp. 71-2; see also Calligaro et al., 2006.
41
Schüssler, Rösch, and Hock, 2001, p. 240.
42
Périn et al., 2007, p. 5.
43
Schüssler, Rösch, and Hock, 2001, p. 241.
44
Bopearachchi, 1999, p. 16.
45
Ibid., p. 13.
 Local Networks and Long-Distance Trade 61
More generally, the compositional analyses of glass beads in Sri
Lanka and in South India deserve attention. Let us compare for
example the results for Tamil Nadu46 and for Sri Lanka.47 In both
areas the mineral lime soda glass (m-Na-Al) dominates with 41 and
43 per cent of the samples respectively: this glass was probably
produced in South India, on sites like Appur or Manikollai, and
largely distributed in South Asia. No production site is known so
far in Sri Lanka: it may thus have been imported from India. How-
ever, the mixed soda-potash glass seems to have been produced on
the site of Giribawa on the west coast of Sri Lanka, where furnaces
and slags have been put to light, as well as raw fragments showing
this particular composition, together with finished beads of m-
Na-Al. Interestingly, it represents 20 per cent of the glass in Sri
Lanka, and 2 per cent for Tamil Nadu:48 it corresponds to orange
annular beads, widely distributed, and red disc-shaped beads, only
found in South India and in Sri Lanka.49 As they represent a big
amount in Kelaniya as well (33 per cent), their production on the
west coast and export to India seems likely. It shows how the west
coast of Sri Lanka could have been in contact with the east coast of
India and with the East in general, supposing the circumnaviga-
tion of the island. Another noticeable point is the presence in
Kelaniya (8 per cent of the glass) and in the southern sites of
Tissamaharama (26 per cent) and Ridiyagama (19 per cent) of the
mineral soda calcareous glass with alumina (m-Na-Ca), which is
often associated with cobalt blue beads of South-East Asian origin,
suggesting as well that the local sailors were used to travel around
the island, probably using the southern sites as redistribution centres
(Tissamaharama showing for example a very balanced picture be-
tween m-Na-Al, mixed soda-potash and m-Na-Ca, each represent-
ing 26 per cent of the analysed glass).
The recent discovery of a shipwreck in the south, close to the

46
Gratuze and Guillaume, 2012.
47
James Lankton, 2013.
48
Gratuze and Guillaume, 2012, p. 140.
49
Ibid., p. 133.
62 Ariane de Saxcé
harbour of Godavaya, requires attention as another indication of
navigation around the island at an early date. The harbour of Goda-
vaya had already given some pieces of evidence for trade, through a
Prakrit-Br"ahm$û inscription of the second century CE50 located in
the Buddhist vihara on top of the rock dominating the sea. It
refers to the grant by the king of the duties of the market place to
the monastery51 and has been interpreted as an indication of the
customs levied in the port on the goods coming from the sea-
trade.52 The monastery could even have been involved somehow in
trade and sailing activities, as the st"upa was a clearly visible land-
mark from the sea for the sailors.
The shipwreck itself ,53 radio-carbon dated of the second or the
first century BCE,54 gives some clues for relations with South India
and, if this is confirmed, will be the best evidence for navigation
along the east coasts of India and Sri Lanka until the southern
ports of the country, and maybe further. The biggest part of the
cargo visible on surface consists of iron, mingled with wooden
pieces,55 and might be related to the iron industry put to light in
Ridiyagama, the city connected to Godavaya through the Walawe
Ganga. As most of the ‘Indian Iron’ mentioned in the classical
texts56 must have included Sri Lankan iron or steel,57 the produc-
tion from Ridiyagama and the cargo of the Godavaya shipwreck
might have been part of a local exchange with India, premise of a
long distance trade later with the Red Sea.
Secondly, one of the stone querns discovered is adorned with three
different auspicious symbols: two fishes, one nandipada and one
« sr∂vatsa. Even though the last two are Brahmanical symbols, the
first related to ®Siva and the second to Vi|s^nu, they are present in a
Buddhist context on the ayaka of the Jetavanarama Dagoba for
50
Paranavitana, 1970, p. 101.
51
Falk, 2001, pp. 327-30, referring to the interpretations of Hettiarachchi,
Bopearachchi, Roth and Kessler.
52
Ibid., p. 329.
53
Bopearachchi, Dissanayaka, and Perera, 2013.
54
Trethewey, 2012, p. 29.
55
Ibid.
56
Periplus 6 for instance.
57
Bopearachchi, 1999, p. 16.
 Local Networks and Long-Distance Trade 63
example.58 Besides, the discovery of similar grinding stones in the
premises of the Yatala st"upa in Tissamaharama invites one to make
the hypothesis of a Buddhist context for the shipwreck symbols.
Joint with the inscription mentioned above, these signs may allow
us to suggest some connection of the ship with Buddhist trading
activities, maybe circulating between the Buddhist sites of India
and Sri Lanka with several stops along the coast.
The last element that hints at a trade link with South India is
the chemical composition of the glass. Some semi-circular glass
ingots have been discovered on the surface of the site and two
samples have been analysed.59 Their chemical composition con-
sists of mineral soda glass with high alumina (m-Na-Al), probably
dating to the second or the first century BCE. The general compo-
sition indicates a possible manufacture in South India, as it shares
many chemical features with the glasses found on the Tamil coast,
where manufacturing sites are known like Appur or Manikollai,
even though the trace elements do not exactly match with any site
in particular, at this stage of our knowledge. Thus it does not cor-
respond to a particular glass from Arikamedu, nor to the Sri Lankan
site of Giribawa, with its furnaces of mixed alcali glass. The Godavaya
glass has a very high percentage of vanadium, which has no exact
equivalent on excavated sites of India or Sri Lanka. The closest one
could be Alagankulam, which is very significant as it is the closest
port to Sri Lanka, and as a whole ‘the results strongly suggest a
South Indian origin for the Godavaya glass’.60
Thus, the Godavaya shipwreck gives another clue for an early
Indo-Sri Lankan trade and for the circumnavigation of the island
by sailors at an early date, long before the Greeks started using this
route, probably by the time of Ptolemy in the second century CE,
as his good knowledge of the island indicates.
Lastly, evidence of traders circulating between Tami|lakam and
Sri Lanka is given by seven Sinhala-Prakrit inscriptions, written in
the Br"ahm$û script, found in Tamil Nadu on the sites of Arikamedu

58
Bopearachchi, Dissanayaka, and Perera, 2013.
59
James Lankton and Bernard Gratuze, in Bopearachchi, Dissanayaka, and
Perera, 2013.
60
Ibid.
64 Ariane de Saxcé
(2), Alagankulam (4) and Kodumanal (1), and listed by I. Maha-
devan.61 They have among other specific linguistic features some
genitive suffixes in -« sa or -ha and the de-aspiration of the aspirates
typical of Sinhala-Prakrit.62 They are dated palaeographically or by
stratigraphy from the second century BCE to the first century CE.
‘The occurrence of these characters in Tamil Nadu, particularly at
the ports and trade centers, suggest their frequent interaction.’63
Vice versa, inscriptions referring to Tamil names in Sri Lanka
testify for the presence of south Indian traders on the island. Thus,
one inscription in Anuradhapura refers to a Tamil householder
while two inscriptions from Periya-Puliyankulam allude to a Tamil
merchant called Vis"a ka.64 An inscribed coin reading mah"a cattan
and alluding probably to a big Tamil trader65 and coins discovered
in the area of Tissamaharama bearing Tamil names66 are other evi-
dence.
This constant interaction between the Indian Peninsula and Sri
Lanka, where specific products such as pearls, textiles or precious
stones were in demand, according to Indian and classical literary
accounts, paved the way for the later intermediary role of Indian
middlemen with western merchants, as the Periplus testifies in its
lists of goods available in the markets of India and Sri Lanka.

THE TRADE PATTERN DURING AND

AFTER THE ROMAN EMPIRE

THE EARLY EMPIRE

By the time of the Periplus in the first century CE, and during the
Early Empire, it seems that these links were used by Indians to
take care of the trade with the West and keep Sri Lanka away from
61
Mahadevan, 1996, two more inscriptions studied in this article come from
West Bengal.
62
Ibid., p. 61.
63
Rajan, 2002, p. 92.
64
Bopearachchi, 2008, pp. 16-17.
65
Rajan, 2002, p. 94.
66
Bopearachchi, 2008, pp. 17-20.
 Local Networks and Long-Distance Trade 65
direct transactions with merchants of the Red Sea. It is broadly
attested by the contrast between thousands of early imperial denarii
and aurei discovered in India and their nearly absence in Sri Lanka.67
But the Periplus itself and some other kinds of evidence also testify
for the non-direct connections of Sri Lanka with the western areas.
Two well-known passages of the Periplus and their interpretation
by Casson68 deserve attention. In the paragraph 60, the author
distinguishes the arrival of goods from the West once a year in Muziris
and a continuous trade on the coasts throughout the year, distrib-
uting slowly the merchandise and money arrived all at once. It
means that local boats were dedicated to the transshipment and
carriage of goods from the west coast to the east coast, hypothesis
reinforced by the distinctions made by the author between three
sorts of boats. The local ones, topika ploia, are doing the coastal
travels, the biggest ones, kolandiophonta, are bound for the East
and China; the third category, sangara, are described as double
monoxylous united by a platform. If there is, as it seems, a special-
ization of craft depending on their destination, the first category
going west, the third one going east, it seems that those double
monoxylous or outrigger boats are going to Sri Lanka. The double
monoxylous are still in use today as ferries but only on rivers, and
would not seem very stable at sea, so that the Greek phrase (liter-
ally ‘big joint monoxylous’) might be interpreted as ships with
outrigger, maybe of the Sri Lankan type yatra dhoni.
As far as networks are concerned, one may attempt a different
kind of exegesis of the Periplus. The analysis of the text content
with tools of social network analysis is interesting in a heuristic
point of view. When creating a diagram of the different categories
of imported and exported goods, Barygaza appears unsurprisingly
in the first position for the diversity of the goods exchanged, thus
obviously confirming its role of transshipment point. It can be
noted that the carnelian referred to earlier is included in the goods
transiting from the Deccan sites to the Indian Ocean through
Barygaza.69 It could have been one of the products traded by the
67
Bopearachchi, 1992.
68
Casson, 1989.
69
Periplus 48, 49.
66 Ariane de Saxcé
sailor mentioned in the Br"ahm$û inscription from Bagavalena, indi-
cating the gift of a cave by this sailor travelling to Bharukaccha.70
The diagram reveals however that the ports of Nelcynda, Muziris
and Bacarê appear also in good position.
The representation through graphs makes it visible that Barygaza
in the first place, and Muziris in the second, consisted in big ex-
change places, importing even more than they exported. On the
contrary, the situation of Taprobanê, only indicated as exporting,
prevents any interpretation as an intermediary market. More in-
terestingly, among the four items mentioned for Taprobanê, three
have a shared network with Muziris: the gems, the pearls and the
tortoise shell; the fourth one—the textiles—is in common with
Barygaza. It points out a specialization in the networks, probably
related to questions of demand and supply: it would have been
more convenient to export gems and pearls to the Malabar coast,
where western merchants came in search of beryl and were prob-
ably specialized in gems and luxury items, and to send the textiles
to Barygaza, where they could find their way with famous muslins
and cotton from Ter or Paithan, for merchants with other specifi-
cations.
If we go further and check on the graph what would be the path
between Taprobanê and Muziris, we find Barygaza as an interme-
diary, which could seem absurd in terms of geography but not
necessarily in terms of economics and can broaden our understand-
ing of the complexity of trade networks. Sri Lanka can be export-
ing textiles to Barygaza, and Muziris importing grain from Barygaza,
specially as grain is mentioned as an import and an export in Bary-
gaza. At the end of the fourth century BCE, Onesicritus already des-
cribed Barygaza as a reloading point, especially for Sri Lanka.71
The situation seems to have continued in the time of the Periplus,
when Mediterranean merchants could take advantage of it.
The gems and other goods could be exported from Sri Lanka to
India and from there to the west, as some sapphires attested in
Italy can testify. The mummy of a young girl, discovered on the

70
Bopearachchi, 2002, p. 107.
71
Schenk, 2007, p. 63.
 Local Networks and Long-Distance Trade 67
site of Grottarossa to the north of Rome,72 buried in a sarcopha-
gus, was not only found with an ivory doll and wearing a dress of
Chinese silk but she also wore a necklace of gold and sapphires,
which were analyzed as originating from Sri Lanka. The type of
burial, the objects accompanying the dead and the decoration of
the sarcophagus, with scenes from the Eneide, indicate a date of
the second century CE. The type of embalming and the physical
condition of the young girl, rachitic and maybe phthisical, hints
at a travel in Egypt, maybe in Alexandria, which climate was re-
puted to cure pulmonary diseases73 and would have been aimed at
healing the little girl, later embalmed so that she could be brought
back to Rome. Egypt, where Alexandria was the nodal point for
the supply of eastern goods, would have been the best place to
find the mortuary offerings. The collection of ivory (probably from
India or Sri Lanka), silk from China and sapphires from Sri Lanka,
gives a good image of the intermediary role of India in the first
centuries of the Common Era, directly linked to Egypt.
The study of pottery from South Arabian and Persian Gulf sites
shows the active role played by Indians in the trade and these ports
as reloading places before going further west. Kanê and Muza are
also referred to in the Periplus as transshipment points and, for
example, Muza is mentioned as importing and exporting ivory
and tortoise shell, both products for which Sri Lanka is one pos-
sible origin, among other possibilities. The study of Turquoise
Glazed pottery from Tissamaharama has evidenced the presence of
sherds with a range of date from the first century BCE to the first
half of the first century CE.74 Their context shows that they have
been discarded in the second century CE. Contrary to what was
usually thought, it reveals contacts between the Persian Gulf and
Sri Lanka before the Sassanids, during the Parthian period. Heidrun
Schenk shows that transshipment centers like Qana and Sumhuram
already worked in the first century BCE passing goods from Meso-
potamia to Sri Lanka, probably via the intermediary of Muziris,

72
Scamuzzi, 1964.
73
Scamuzzi, 1965, 77-80
74
Schenk, 2007, p. 62.
68 Ariane de Saxcé
where 1,736 fragments of Turquoise Glazed pottery and 3,684
sherds of torpedo jar from Mesopotamia have been discovered during
the seven seasons of excavations.75 The latter type reveals the con-
tinuation of trade by the end of the Roman Empire and later.

THE LATE EMPIRE AND AFTER

From the fourth and fifth centuries CE onwards, it is well known
that the links of Sri Lanka with Rome and with Persia appear to
have taken a direct route, made obvious by the thousands of Roman
third brass found on the island, that seem to have been in circula-
tion, as the example of Sigiriya, where no other means of exchange
has been found, testifies.76
As much as Romans, Persians and South Arabians seem to have
played a dominant role in the trade during this period. In Sigiriya,
Osmund Bopearachchi has put into relief the importance of the
model of the Moghol gardens in the conception and the organiza-
tion of the Sigiriya pleasure gardens.77 He has shown how the
symmetrical manipulation of water, the frequent presence of stone
water channels and the fourfold division of the garden were related
to the Sassanian gardens, themselves influenced by the quadripartite
Achaemenid gardens of Cyrus in Pasargadae.78 Besides, the nestorian
cross found in Anuradhapura can testify of those contacts with
Persia.
The contacts with Mesopotamia are attested through the pres-
ence of ‘torpedo jar’, this late neckless cylindrical jar that is gener-
ally considered as a wine container during Sassanian times.79 It has
been noticed on the island on the sites of Mantai, Anuradhapura
and Tissamaharama,80 plus three sherds on the recently excavated

75
Cherian, 2013.
76
Bopearachchi, 1990.
77
Bopearachchi, 2002, p. 111-13.
78
Bopearachchi, 2002, p. 112.
79
Tomber, 2009, p. 50.
80
Tomber, 2008, p. 126-7.
 Local Networks and Long-Distance Trade 69
site of Kuchchaveli on the East coast.81 The discoveries on that site,
which most important period of occupation is beyond the sixth
century CE, confirm the account of Cosmas that Sri Lanka is a
crossroads for international trade, through the presence of Roman
coins, Red Polished ware, Torpedo jar and Chinese ware altogether,
showing its intermediary place between east and west. The devel-
opment of the studies of Red Polished ware have particularly re-
newed the understanding of the trade links with India. It is not
any more considered only as a sign of first century CE contacts with
Gujarat, as it used to be by reference to Deccan sites and links with
the Roman trade. Careful analyses of typology and paste have shown
late productions of this ware. In Anuradhapura, 92 sherds have
been found, with a date range of 200 BCE to 1100 CE;82 in Tissamaha-
rama, different vessel forms are found, from the first century BCE
layers to the eighth-ninth centuries CE (phase h). It then denotes
early medieval contacts with India, as well as on South Arabian
sites like Qal’at al-Bahrain or al-Shihr (Yémen),83 were this pottery
has been identified, also from a late context. D. Kennet diagnosed
a continuous use as late as the eighth century CE.84 It is interesting
to note that it is then an evidence of late contacts between India,
Sri Lanka and the Arabian peninsula.
Late commercial contacts between Sri Lanka and the West, in
parallel with Indian products, were recently proved also with studies
on garnet jewelry during the Merovingian times in Gaul.85 We
have already alluded to the difference between the Indian and Sri
Lankan garnet. On the cloisonné style, five groups of different gar-
nets have been used. Among them, three are located in South Asia
and are big enough to be employed in big pieces to make those
jewels: one source is in Rajasthan, the type 1 of almandines, another
81
Excavations of the Department of Archaeology, Sri Lanka, and the École
Normale Supérieure, Paris, under the direction of N. Perera and O. Bopearachchi;
publication forthcoming.
82
Coningham, Ford et al. 2006, p. 168.
83
Schenk, 2007, p. 67.
84
Ibid.
85
Calligaro et al., 2006.
70 Ariane de Saxcé
one in Sri Lanka, the type 3 of pyraldines and a third one, the
type 2, of almandines, is not yet identified and may be from South
India or from Rajasthan as well. The type 3 has recently been
divided between Sri Lanka and Orissa. They are the sources of the
major part of the jewelry in France in the sixth century CE and the
cloisonné style stops at the beginning of the seventh century for
unexplained reasons of a shortage in the supplies. This shortage is
made clear by some tombs decorated with jewels prepared for a
cloisonné but without the garnets that are supposed to be inserted.
Nevertheless, it indicates a continuity of trade between Sri Lanka,
India and the West at a late period.

Thus, this short panorama of Sri Lankan participation in the trade

shows that the existant local exchanges established before the di-
rect contacts with the Greco-Roman world have been dense and
rich enough to prepare the way to long distance trade and facili-
tate further transactions. The role of Sri Lanka in the western part
of the Indian Ocean follows an evolving pattern from close rela-
tionships with India playing an intermediary role to a more inter-
national and direct involvement in these exchanges as a trade plat-
form between Rome, China, Persia, Arabia and even the Barbarian
kingdoms after the end of the Roman Empire.

REFERENCES

Allchin, B. 1979, ‘The Agate and Carnelian Industry of Western India and
Pakistan’, in South Asian Archaeology 1975, 91-105, Leiden: Brill.
Bopearachchi, Osmund, 1990, ‘Some Observations on Roman Coins Found in
Recent Excavations at Sigiriya’, Ancient Ceylon 7: 20-37.
——, 1992, ‘Le Commerce Maritime entre Rome et Sri Lanka d’après les
données numismatiques’, Revue des Études Anciennes 94 (1-2): 107-21.
——, 1999, ‘Sites Portuaires et Emporia de l’ancien Sri Lanka: nouvelles données
archéologiques’, Arts Asiatiques, no. 54: 5-22.
——, 2002, ‘Archaeological Evidence on Shipping Communities of Sri Lanka’,
in Ships and the Development of Maritime Technology in the Indian Ocean,
92-127, London; New York: Routledge Curzon.
 Local Networks and Long-Distance Trade 71
——, 2006, ‘Coins’, in Anuradhapura, 7-26. Oxford, Royaume-Uni: Archaeo-
press.
——, 2008, Tamil Traders in Sri Lanka and Sinhalese Traders in Tamil Nadu.
Colombo: International Centre for Ethnic Studies.
Bopearachchi, Osmund, Senarath Dissanayaka and Nimal Perera, 2013,
‘Recent Archaeological Evidence on the Maritime Trade in the Indian
Ocean: Shipwreck at Godavaya’, in The Ports of the Indian Ocean from the
Red Sea to the Gulf of Bengal, Calcutta.
Bopearachchi, Osmund and K. Rajan, 2002, ‘Graffiti Marks of Kodumanal
(India) and Ridiyagama (Sri Lanka)—A Comparative Study’, Man and
Environment XXVII (2): 97-105.
Calligaro, Thomas, Patrick Périn, Françoise Vallet and Jean-Paul Poirot, 2006,
‘Contribution à l’étude des grenats mérovingiens, nouvelles analyses
gemmologiques et géochimiques effectuées au Centre de Recherche et de
Restauration des Musées de France’, Antiquités Nationales 38: 111-44.
Carswell, John, 1991, ‘The Port of Mantai, Sri Lanka’, in Rome and India: The
Ancient Sea Trade, 197-203. Madison (Wis.): University of Wisconsin
press.
Casson, Lionel, 1989, The Periplus Maris Erythraei, Princeton: Princeton
University Press.
Cherian, P. J., 2013, ‘Indian Ocean Maritime Network and the Archaeological
Evidence from Pattanam, India’. Communication in the Seminar Imperial
Rome, Indian Ocean Regions and Muziris. Colloquium held in Irinjalakuda,
Kerala.
Coningham, Robin, 2002, ‘Beyond and Before the Imperial Frontiers: Early
Historic Sri Lanka and the Origins of Indian Ocean Trade’, Man and
Environment XXVII (1): 99-108.
Coningham, Robin, Daniella Burroni, Randolph Donahue, and Louise Ford,
2006, ‘Stone Objects’, in Anuradhapura: The British-Sri Lankan Excava-
tions at Anuradhapura Salgaha Watta 2. Volume 2, The Artefacts: 377-429.
Oxford: Archaeopress.
Coningham, Robin, Louise Ford, Steven Cheshire and Ruth Young, 2006,
‘Unglazed Ceramics’, in Anuradhapura: The British-Sri Lankan Excava-
tions at Anuradhapura Salgaha Watta 2. Volume 2, The Artefacts: 127-331.
Oxford: Archaeopress.
Falk, Harry, 2001, ‘Three Epigraphs from Go?avâya, Sri Lanka’, in Ancient
Ruhuna?: Sri Lankan-German Archaeological Project in the Southern Prov-
ince, 327-34. Mainz: P. von Zabern.
Ford, L. A., A. M. Pollard, R. A. E. Coningham and B. Stern, 2005, ‘A Geochemi-
cal Investigation of the Origin of Rouletted and Other Related South Asian
Fine Wares’, Antiquity 79 (306): 909-20.
72 Ariane de Saxcé
Gogte, Vishwas D., 1997, ‘The Chandraketugarh-Tamluk Region of Bengal:
Source of the Early Historic Rouletted Ware from India and Southeast
Asia’, Man and Environment XXII (1): 69-85.
——, 2001, ‘XRD Analysis of the Rouletted Ware and Other Fine Grey Ware
from Tissamaharama’, in Ancient Ruhuna?: Sri Lankan-German Archaeo-
logical Project in the Southern Province, 197-202, Mainz: P. von Zabern.
Gratuze, Bernard, and Sarah Guillaume, 2012, ‘Analysis of Tamil Nadu Glass
Beads: Application to the Study of Inland Glass Trade’, in New Dimensions
in Tamil Epigraphy, 129-43, Chennai: Cre-A.
Lankton, James W., 2013, ‘Scientific Study of the Godavaya Glass Ingots: New
Insight into Early Glass Production and Exchange in South Asia’, in Mari-
time Trade and Cultural Interactions in the Indian Ocean. Colloquium held
in Sigiriya, Sri Lanka.
Magee, Peter, 2010, ‘Revisiting Indian Rouletted Ware and the Impact of
Indian Ocean Trade in Early Historic South Asia’, Antiquity 84: 1043-54.
Mahadevan, Iravatham, 1996, ‘Old Sinhalese Inscriptions from Indian Ports:
New Evidence for Ancient India-Sri Lanka Contacts’, Journal of the Insti-
tute of Asian Studies XIV: 55-65.
Paranavitana, S. 1970, Inscriptions of Ceylon, Colombo: Archaeological Survey
of Ceylon.
Périn, Patrick, Thomas Calligaro, Françoise Vallet, Jean-Paul Poirot, and
Dominique Bagault, 2007, ‘Provenancing Merovingian Garnets by PIXE
and Ì-Raman Spectrometry’, in Post-Roman Towns, Trade and Settlement in
Europe and Byzantium, 69-75, pl. 5, 6. Berlin; New York: De Gruyter.
Rajan, K. 2002, ‘Maritime Trade in Early Historic Tamil Nadu’, Man and
Environment XXVII (1): 83-98.
Roux, Valentine and Jacques Pelegrin, 2000, Cornaline de l’Inde?: des pratiques
techniques de Cambay aux techno-systeÌmes de l’Indus, Paris: Ed. de la Maison
des sciences de l’homme.
Scamuzzi, Ugo, 1964, ‘Studio Sulla Mummia Di Bambina, Cosidetta “Mummia
Di Grottarossa”, Rinvenuta a Roma, Sulla Via Cassia, Il 5-2-1964’, Rivista
Di Studi Classici XII (3): 264-80.
——, 1965, ‘La ‘Mummia Di Grottarossa’, Aegyptus 45 (1/2): 74-83.
Schenk, Heidrun, 2001, ‘The Development of Pottery at Tissamaharama’, in
Ancient Ruhuna?: Sri Lankan-German Archaeological Project in the Southern
Province, 59-195, Mainz: P. von Zabern.
——, 2006, ‘The Dating and Historical Value of Rouletted Ware’, Zeitschrift
Für Archäologie Außereuropäischer Kulturen 1: 123-52.
——, 2007, ‘Parthian Glazed Pottery from Sri Lanka and the Indian Ocean
Trade’, Zeitschrift Für Archäologie Außereuropäischer Kulturen 2: 57-90.
 Local Networks and Long-Distance Trade 73
Schüssler, U., C. Rösch and R. Hock, 2001, ‘Beads from Ancient Sri Lanka-
First Results of a Systematic Material Analysis’, in Ancient Ruhuna?: Sri
Lankan-German Archaeological Project in the Southern Province, 227-42,
Mainz: P. von Zabern.
Tomber, Roberta, 2008, Indo-Roman Trade?: From Pots to Pepper, London:
Duckworth.
——, 2009, ‘Beyond Western India: The Evidence from Imported Amphorae’,
in Migrations, Trade and Peoples: European Association of South Asian
Archaeologists. Proceedings of the Eighteenth Congress, London, 2005: 42-57,
London: The British Association for South Asian Studies, The British
Academy.
Trethewey, Ken, 2012, ‘The Oldest Known Ship in the Indian Ocean’, INA
Quarterly 39 (1-2): 28-9.
Unesco, 1993, The Cultural Triangle of Sri Lanka, Paris; Colombo: Unesco;
Central Cultural Fund.
Von Rosen, Lissie, 1988, Lapis Lazuli in Geological Contexts and in Ancient
Written Sources. Partille: P. Aström.
Weerakkody, D.P.M., 1997, Taprobanê: Ancient Sri Lanka as Known to Greeks
and Romans, Turnhout: Brepols.