Nikola Theodossiev
Celtic Settlement in North-Western Thrace during the Late Fourth and Third
Centuries BC: Some Historical and Archaeological Notes*
North-western Thrace is a specific region of the northern Balkans that was
inhabited by various tribal communities of Thracians, Illyrians and Celts who
played important roles in the history and culture of south-eastern Europe during the
first millennium BC. Geographically, the ancient region of Thrace covers modern
north-western Bulgaria and eastern Serbia, while the most powerful tribal
community that emerged in the fifth century BC was known as the Triballi who
were Thracian people well described in various Greek and Roman historical
sources (for a detailed study on the whole region see: Theodossiev 2000). The early
Celtic settlement in this part of ancient Thrace has been always a controversial
question examined by different scholars since the first decades of the twentieth
century. Some recent detailed studies have thrown additional light on the whole
problem (Megaw et al. 2000; Megaw 2004; Theodossiev 2000: 82-100) and the
following paper is offered in honour of Zenon Woźniak, an international scholar of
the highest reputation who has been for many years deeply involved in European
Iron Age studies and who is the author of a number of seminal studies of the
eastern zone of La Tène culture (Woźniak 1974; 1976).
Various historical records and archaeological finds of the early Hellenistic
Age provide us with options for considering the presence of Celts in north-western
Thrace and their interaction with local tribal communities. The earliest contacts are
2
evidenced in Arrianus (Anab. I 4, 6-8) and Strabo (VII 3, 8) who testify that Celts
from the Adriatic coast attended the armistice and alliance agreements concluded
between the Triballian king Syrmos and Alexander the Great in 335 BC (see also
Bouzek and Guštin in this volume). During the negotiations, which took place in
the territory of the Triballi, the Celts affirmed their vows of friendship and
hospitality before the Macedonian king and presumably at precisely that time they
established first diplomatic contacts with Triballi (for a compilation of the ancient
records see: Katsarov 1919; Cunliffe 1997: 79-80; Megaw 2004; Theodossiev
2000: 81-82; see also discussion of the nature of ancient literary sources concerning
Celts in: Dobesch 1991; Rankin 1995).
Any reconstruction of contacts and relations, including gifts exchange,
between Celts and Triballi during the last decades of the fourth and the very
beginning of the third century BC, or, in contrast, supposed military conflicts and
invasion, does find some support in the scanty archaeological finds available. Such
is the golden Celtic neckring or torc (Fig. 1), unearthed by chance near Gorni
Tsibar, a village situated close to the southern bank of the Danube in north-western
Bulgaria (Theodossiev 2000: 116, cat. no. 84 with op. cit., fig. 90; cf Wells 1995
for the possible range of methods of distribution of valuable objects). The torc dates
to the end of the fourth or beginning of the third century BC and finds good
parallels among a number of La Tène B1/B2 neckrings discovered in other parts of
Iron Age Europe (Jacobsthal 1969: 170, no. 46; Megaw and Megaw 2001: 119
ill.168; Moscati et al. 1991: 712: no. 129; more recently see also: Megaw 2004). It
is known that the torc was considered as a sacred object and symbol of high social
status among the Celts (Green 1992: 211-212). Unfortunately, as far as the Gorni
Tsibar neckring is concerned, there is no reliable information on the find
circumstances or its archaeological context. Therefore, it is only speculation to
consider that the torc originally formed part of a Celtic burial or ritual deposit and
3
thus could be related to a presumed ethnic presence of Celts in that region of
Thrace. However, such a presence during the late fourth and early third century BC
is not supported by any surviving historical record. Since the torc may alternatively
come from a Thracian funeral or other ritual context, it is possible to assume that
the precious object was a political gift, offered by some Gaulish chieftain to an
unknown Triballian aristocrat during the course of negotiations. One may go
further and hypothesise that similar contacts accompanied the first expansion of
Celtic tribal groups towards the central Balkans during the last decades of the
fourth century BC, as well as during their settlement on the western fringes of
north-western Thrace, in the Morava river valley, previously controlled by the
Illyrian tribe of the Autariatai. Of course, the Gorni Tsibar torc might be also
interpreted as booty, related to the victory of Cassander over the Galatae, which
took place in the area of the west Balkan range c. 310 or c. 298 BC (Frey, Szabó
1991: 481 who suggest a relationship between the torc and this historical event).
From the first decades of the third century BC the territory of the Triballi
became a marginal zone of the Hellenistic world and during the late 280s and 270s
BC indigenous tribal communities faced a well-organized Celtic military invasion
(on this incursion and the political history of the region see: Katsarov 1919;
Polaschek 1937; Mócsy 1966; Gerov 1967; 1969 with historical sources; Danov
1975-1976; Papazoglu 1978: 52-57, 272-278; 1988; Domaradski 1984; Tacheva
1987: 27-48; Hammond 1989: 298-301; Frey, Szabó 1991; Garašanin 1996;
Lazarov 1996; Theodossiev 2000: 82-92; Megaw 2004). However, the great Celtic
inrush into north-western Thrace was preceded by military pressure on its western
periphery and incursions leading towards the central Balkans, incursions that had
already begun in the last decades of the fourth century BC, when, c. 313 BC, the
Gauls conquered the Illyrian Autariatae and banished part of them. Soon after that
conquest, in 310 BC — or possibly 298 BC — the Macedonian general Cassander
4
defeated the Celts somewhere in the west Balkan range, presumably in the territory
of the Triballi as mentioned above. Supposedly at the same time, military
detachments of Gauls, led by Cambaules, reached the borders of Thrace but did not
have the courage to invade it.
While the gold torc from Gorni Tsibar may not testify with complete
certainty to military conflicts or to a Celtic enclave in north-western Thrace during
the late fourth or beginning of the third century BC, the conquest of the Autariatae
c. 313 BC and Celtic settlement in the Morava river valley are well evidenced with
archaeological material. Thus, the investigation of the flat cemetery in the region of
Pecine near Kostolac, a town situated in the lower Morava valley and close to the
Danube, in north-eastern Serbia, provides a good illustration of the ethnic changes
that occurred in the region (Jovanović 1984, 1985, 1991; Theodossiev 2000: 120-
121, cat. no. 113 with full bibliography). A number of Celtic cremation and
inhumation graves, the earliest dating to the end of the fourth or beginning of the
third centuries BC, are situated around the nine earliest graves, presumably
belonging to the Autariatae and dated to the second half of the fourth century BC.
These are located in the centre of the cemetery (on possible ethnic identification of
the earliest burials see: Theodossiev 2000: 40-41). The continuity observable on
this burial site clearly indicates that the new Celtic settlers did not annihilate the
Autariataec community but that they assimilated with the indigenous population
and presumably mixed ethnically with it (Jovanović 1985, 1992). Therefore, it is
possible to assume that from the end of the fourth century BC onwards, the Morava
River valley and the regions located to the east turned into a Celto-Illyro-Thracian
interaction zone (Theodossiev 2000: 98-100).
The major and well-organized Celtic military invasion into Thrace,
Macedonia and Greece occurred at the very end of the 280s BC. Using as a base the
already conquered territories of the Autariatae in the central Balkans, in 280 BC the
5
Gauls led by Cerethrius prepared themselves to fight against the Triballi and the
remaining Thracians, Brennus and Acichorius against Paeonia, and Bolgios/Bulgius
against Macedonia and Illyria (Pausanias X 19, 7). At the end of the same year or in
the next one, Bolgius defeated the Macedonian army and killed Ptolemy Ceraunos.
In 279 BC, Brennus— possibly not a historical character — began a campaign
against Delphi; the Celtic detachments on their way south devastated the Dardani,
neighbours to the south-west of the Triballi. At the same time, those Gauls who
were left by Brennus to guard their tribal borders, enlisted 3000 horsemen and 15
000 foot-soldiers to form an army which defeated the Triballi and the Getae, who
seem unsuccessfully to have tried to push out the invaders (Iustinus XXV 1, 2-3).
In the same year, 279 BC, after the complete defeat of the Celts during the
campaign against Delphi and Brennus’s suicide, some of the surviving Gauls took
the way back led by Bathanattus (later the route was named the ‘Way of
Bathanattus’) and in 278 BC part of them, the so-called Scordisci, settled the lands
between Sava, Danube and Morava rivers (Polybius I 6, 4; Iustinus XXIV 4, 8,
XXXII 3, 6-8; Posidonius frg. 48J; Athenaeus VI 25, p. 234 a-b). The new
Scordiscan settlers completely overran the local Autariatae and formed a powerful
tribal alliance that played an important rôle, both politically and military, in the
north Balkans until the end of the first century BC (Garašanin 1966, 1996;
Todorović 1974; Papazoglu 1978; Jovanović, Popović 1991). Also in 279 BC,
another part of the Celtic forces, who survived the defeat at Delphi, led by
Comontorios established a tribal state in Thrace with its capital called Tylis,
presumably located in the region of Byzantium, where another group of Gauls
under the leadership of Leonorius and Loutorius had already settled after they had
separated from Brennus during the march through Dardania. The Gaulish tribal
state established in this way expanded gradually and reached the Balkan range to
the north but was destroyed by the Thracians in 213 BC (recent study in: Lazarov
6
1996). In 277 BC, Antigonus Gonatas defeated the Gaulish army in a great battle
near Lysimacheia, close to the Thracian Chersonese. Thus, he ended the period of
Celtic invasion and plundering.
The early settlement of the Scordisci on the western periphery of northwestern
Thrace, that is, in the lands to the east of the Morava, is not well
documented in the ancient sources. Strabo (VII 5, 11) provides vague information
on the Triballi being conquered by the Autariatae, and one possible interpretation of
this text is that both tribal groups were defeated by the newly settled Scordisci in
278 BC (Papazoglu 1978: 53). This interpretation finds some support in Appianus
(Illyr. 3) who notes that after the Scordisci defeated the Triballi, the latter withdrew
beyond the Danube towards the territory of the Getae. Strabo (VII 3, 13) also
describes similar movements of the Getae to the south of the Danube, due to the
military pressure from the Scythians, Bastarnae and Sauromatae, while on the other
hand, the Triballi, threatened by Illyrians (most likely the Scordisci: Papazoglu
1978: 54-55), escaped to the north of Danube. Unfortunately, the ancient records
cannot be placed in a reliable chronological framework (Papazoglu 1978: 54). Most
likely, these military confrontations occurred in the 270s and 260s BC, not long
after the Triballi had been defeated for the first time by the Brennus’s Celts in 279
BC, and when the Scordisci led by Bathanattus had already settled between the
Morava and Sava rivers (Garašanin 1996). The Triballian tribes, who had escaped
across the Danube, presumably inhabited the region immediately to the east of the
Morava, where the Little Scordisci were certainly located during the second and
first centuries BC. During the second quarter of the third century BC the Triballian
tribal community, after it had experienced the effects of the devastating Celtic
invasion, certainly lost forever political control of the peripheral western area.
A number of archaeological finds may be interpreted as a proof of the fluid
political situation within north-western Thrace due to the Gaulish invasion and the
7
following conflicts with the newly settled Scordisci. In primary position is the
remarkable silver treasure from Rogozen, which most likely was hidden as a result
of military threat. If so, the latest vessels, dated to the end of the fourth century and
first decades of the third century BC, testify that the hoard was buried in the period
of the great Celtic invasion of 280 BC (Theodossiev 2000: 135, cat. no. 196).
About eleven separate hoards, consisting of early Hellenistic gold and silver coins,
are known from north-western Thrace and presumably most of these were also
hidden during the turbulent events of 270s and 260s BC when the Triballian tribes
had been defeated by the Gauls and the Scordisci (Theodossiev 2000: 84, 101 and
List).
The ancient sources describing Celtic incursions are well supported by other
archaeological finds. In 1959, an iron La Tène C1 sword with fragmentary
decorated scabbard (Fig. 2a, b) was found in a funerary pyre, during excavation of
the Kopanata Mogila tumulus at c. 2 km to the north of Pavolche, a village situated
near the town of Vratsa in north-western Bulgaria (Nikolov 1965: 179-181;
Theodossiev 2000: 143, cat. no. 248; Megaw et al. 2000). The tumulus belongs to a
barrow cemetery consisting of six small mounds located close to the South of the
Vrats-Mezdra road. Although the decorated sword scabbard has been previously
considered in a number of prior publications (Woźniak 1974: 46-95, fig. 9: 6, 190,
list II, no. 109), only recently has a detailed study provided a complete analysis of
the find and in particular the decorated scabbard. The scabbard design, which most
probably is derived from the dragon- or bird-pair motifs, may be compared with a
number of decorated scabbards and swords known in Slovenia, Croatia and
Transdanubia, a style which spread throughout Iron Age Europe in La Tène B and
C; undoubtedly, the Pavolche sword is a Celtic product (Megaw et al. 2000;
Megaw 2004). Rather than an occasional import or status-enhancing gift or booty
belonging to a deceased Triballian warrior, it might be suggested that the Pavolche
8
sword might be evidence of a Scordiscan intruder buried in the mound some time
around 270-200 BC. However, the funerary rite and grave construction — a pyre
within a tumulus — evidenced by the Pavolche barrows is entirely typical of the
Triballi during the Hellenistic period and is not known among the Scordisci
(Theodossiev 2000: 29-31, 44-45); moreover, almost all grave goods found at
Pavolche are of Thracian origin. Therefore, having in mind that there is not a single
historical record testifying to Celtic settlement in the eastern part of north-western
Thrace during the third century BC, and given that all sources describe a period of
invasion and conflict, and moreover, that at present there are no excavated graves in
the region that can be regarded as typically Celtic (cf Theodossiev 2000: 25-48),
one may conclude that the Pavolche sword was most likely booty or a gift buried
within a Triballian funerary context. Of course, one may not exclude the probability
that some Gaulish (that is, Scordiscan) ethnic enclaves existed in the region during
the third century BC (cf Zirra 1976; Tacheva-Hitova 1978; for Celtic enclaves in
north-eastern Thrace: Lazarov 1996). Following from this, any Celts who may have
settled among the indigenous Triballi would have been assimilated and adopted
Thracian funerary customs. In any case, the Pavolche sword scabbard reflects the
complex and often violent interactions and the dynamic relationship between Celts
and Thracians in a period of war, invasion and ethnic changes.
Another curious object, which may suggest contacts and interaction between
the Scordisci and Triballi, is a small inscribed cult relief (Fig. 3a-c), a chance find
most likely from western Bulgaria and currently located in the National Museum of
Archaeology, Sofia (Manov 1993; Theodossiev 2000: 58, fig. 155). The carved
stone illustrates well the religious beliefs of the Scordisci who obviously worshiped
Epona, the tribal ancestor-god and the warrior hero. Since the object comes from no
clear archaeological context and lacks exact parallels among Celtic cult reliefs, it is
not easy to provide a precise dating for the piece; only the inscription gives some
9
support for a third century BC date (Manov 1993). The ritual function of the object
is also unknown but carved stones displaying various imagery were widely used in
Celtic cult practices (Cunliffe 1992; 1997; Green 1986; 1992; 1995: 466-468;
Megaw and Megaw 2001; Moscati et al. 1991). While the ritual purpose is obscure,
the understanding of the figures carved on the Sofia object seems to be relatively
clear. On the one side of the relief, there is a mare, which in this particular case
might be interpreted as a hippomorphic personification of Epona, the Celtic horsegoddess
(Green 1986: 91-94, 173-174; 1992: 90-92; 1995: 479). Epona was known
as a deity of fertility and prosperity but she was also associated with beliefs
relevant to death and the underworld. The other side of the carved stone shows a
man in a fight to the death with an enormous snake. Most probably, this is a
representation of the tribal warrior-chieftain and hero, presumably regarded as a
mythical ancestor and represented in a moment of heroic confrontation, fighting a
chthonic reptile (cf. Green 1986: 185-186; 1992: 194-195 on the healing, chthonic,
fertility, protection and evil associations of the snake in the Celtic cults). In Irish
mythology, the warrior chief Finn mac Cumhaill kills great water snakes, while the
Ulster hero Conall Cernach is recorded as overcoming an enormous snake which
guards a treasure (Green 1992: 64-65, 98-99, 194-195). On the edge of the Sofia
relief, a second snake is depicted together with a male facing to the front
contiguous to a short incised inscription – ΣΚΟΡΔΟ (= genitive: ‘belonging to
Scordus’). It is beyond any doubt that this is the image of the tribal eponym and
ancestor-god Scordus, attested as Scordiscus in the sources (Appianus, Illyr. 2).
Although the fighting scene between the Scordiscan warrior hero and the snake
finds good parallels in later Insular Celtic myths, an alternative reading would be
that the imagery was influenced by the Delphic dragon-slaying myth concerning the
fight between Apollo and Python. In such a case, it is possible even to assume a
Graeco-Celtic religious syncretism and to suggest that the carved cult stone was
10
produced after the failed Gaulish campaign against Delphi in 279 BC. As discussed
above, this event was of crucial importance for the establishment of the Scordiscan
tribal community and its later political development. It is interesting that the cult
relief is the first monument, which displays the use of Greek script by Scordisci in
the Hellenistic period.
A number of historical records of the late second and first centuries BC reinforce
the earlier evidence on multilateral interaction and relationship, clearly testifying
that the Scordisci inhabited the western regions of north-western Thrace and were
there intermingled with Thracians and Illyrians. Ammianus Marcellinus (XXVII 4,
4) mentions that the Scordisci occupied part of Thrace while Strabo (VII 5, 12)
specifies that the Great Scordisci lived between the Noaros (present-day Sava or
Drava) and the Morava rivers, while the Little Scordisci inhabited lands to the east
of the Morava in the neighborhood of the Triballi and Moesi (for the distribution of
the Scordisci in the Hellenistic age see: Papazoglu 1978: 354-389). Moreover, both
archaeological finds and historical sources testify that during the late Hellenistic
period the Little Scordisci became ethnically intermingled with the indigenous
Autariatae and Triballian tribes in the Morava valley and further to the east; this led
to syncretic ethnic and cultural processes and turned the region into a Celto-Illyro-
Thracian interaction zone (Garašanin 1957, 1966; Todorović 1966; Zirra 1976;
Jovanović 1992; Theodossiev 2000: 85-92, 98-100; Gerov 1967). The co-existence
of different tribes and their ethnic mixing are clearly evidenced by Strabo (VII 3, 2
et 11; VII 5, 1-2), who describes Scordisci crossed with Thracians and Illyrians and
writes of Thracians mixed with Celts and Scythians, obviously having in mind
north-eastern Thrace and the regions to the north of the Danube. Presumably, it is
due to these reasons that other ancient authors, when they describe Celtic ferocity,
also identify the Scordisci with the Thracians, (Iulius Florus, Epitom. de T. Liv. I 39
et III 4; Rufius Festus IX; Orosius, Hist. adv. pag. V 23, 17-19; Iordanes, Rom.
11
219). As discussed above, the Celtic flat cemetery at Kostolac clearly indicates
ethnic change and assimilation, which occurred at the end of the fourth century and
into the first decades of the third centuries BC, before the region fell under the
political control of the Scordisci during the 270s BC. Therefore, it is possible to
suppose that the historical sources, describing the localization of the Little
Scordisci and the ethnic mixing between Celts, Thracians and Illyrians during the
late second and first centuries BC, may reflect earlier events at the time of the great
Gaulish invasion, which was to have so strong an impact and to change for ever
certain parts of the early Hellenistic world.
Dr. Nikola Theodossiev
Department of Archaeology
Sofia University St. Kliment Ohridski
Bul. Tsar Osvoboditel 15
Sofia 1504, Bulgaria
