Byzantine Dress Accessories in North Africa:
Koiné and Regionality
Christoph Eger
Byzantine dress accessories in North Africa: an ambiguous
term
‘Byzantine dress accessories in North Africa’ has a twofold
meaning: either dress accessories of Byzantine origin or of the
Byzantine period. The two are not necessarily identical. For
more than 160 years, from 533/4 until 698, North Africa was
ruled by Byzantium and produced one of the most important
emperors of the time, Heraclius (610–41). However, the region
between Septem (mod. Ceuta) and the border to Cyrenaica was
not one of the core regions of the Byzantine Empire, but was
situated on its south-western periphery. And it did not belong
to the Byzantine world from the outset: the Late Antique
period in North Africa can be divided into two phases: until
429 the whole of North Africa formed part of the western half
of the Roman Empire; then from 429 until 533/4 the Vandals
controlled major parts of the former diocese of Africa. The
geographical situation and historical development alone make
quite clear that the material culture cannot have been shaped
solely by eastern Mediterranean influences, but was also
influenced from elsewhere.
In North Africa, as in other regions of the Byzantine
Empire, supra-regional types and fashions sometimes
co-existed with local traditions. In terms of provincial
Byzantine archaeology, it is therefore desirable to establish
a clearer regional differentiation of the find material which
records the idiosyncracies of craftwork within the Byzantine
Empire. Despite the still unsatisfactory state of publication this
seems possible, at least for some groups of objects, amongst
them belt buckles.1 The pre-condition for this is a detailed
typological and stylistic analysis, which also includes technical
details as far as possible. On this basis, types and variants
can be defined and their distribution patterns consequently
analysed. It has long been known that the main distribution
concentrations of Byzantine small finds do not necessarily
indicate major market areas where production sites may be
assumed. Inevitable filters for the interpretation of distribution
maps are the different quality and quantity of available
source material (e.g. different burial customs, which become
particularly apparent in finds outside the Empire) and the
widely variable state of publication in the various regions of the
Mediterranean. The last is particularly true for North Africa
where, hitherto, it has been impossible to obtain a clear idea of
Late Antique dress accessories.
Source criticism
On many distribution maps of Late Antique metal small finds,
North Africa still appears mainly as a blank. Apart from the
corpus-like publication of bronze finds from Morocco,2 there
is a lack of comprehensive find-spot analyses.3 Finds of the
Vandal period, which have long received special attention from
Early Medieval archaeologists and which have been intensively
discussed recently,4 are comparatively well known. Compared
with these studies, the state of publication of small finds of the
Byzantine period has to be considered very poor. Essentially,
one can only refer to a few pieces from Hippo Regius (mod.
Annaba) published by Marec in 1958,5 and the finds from the
recent excavations at Carthage supported by UNESCO. The
final excavation reports of the British, American and German
excavations deserve particular mention as they also included
the small finds. In particular the German excavations in the
coastal sector known as the ‘Quartier Magon’ unearthed a
considerable spectrum of finds of Late Antique and Byzantine
objects considering its relatively small surface area.6
Almost unknown are finds deriving from the excavations
and collections of the White Fathers (Pères blancs, today:
Missionari d’Africa) in Carthage over many decades. Their
activities are inseparably linked with the name of Alfred-Louis
Delattre, who arrived in Carthage in 1875 as a young monk and
started the first excavations in the area of the basilica extra
muros at Damous al-Karita only a few years later. For more
than 50 years he actively participated in the archaeological
investigation of ancient Carthage and created an impressive
collection of material not only from his own excavations, but
also by buying stray finds from Carthage and its environs. He
made them accessible to the public in the Musée Lavigerie de
Saint-Louis on the Byrsa hill. This collection formed the basis
of the present-day Musée National de Carthage. Unfortunately,
and quite unexpectedly given his long publication list, Delattre
never wrote a line on many groups of objects including Late
Antique fibulae and belt buckles.7 This may be the reason why
one of the largest collections of Late Antique dress accessories
of the Mediterranean has escaped scholarly attention for such
a long time. The present inventory of the Musée National de
Carthage includes around a dozen fibulae and more than
200 buckles and buckle fragments of the 5th to 7th centuries,
which I have studied in the last few years as a research project
devoted to Late Antique dress accessories from North Africa.8
Finds from eastern Algeria and western Libya, amongst
them objects from Thamugadi (mod. Timgad), Cuicul (mod.
Jamilah) and Sabratha (mod. Sabrāta), complete the find-spot
spectrum. All three cities have been investigated over large
areas: at Timgad, for instance, nearly all the urban areas plus
a Byzantine fortress have been excavated. Mention should
also be made of the huge southern necropolis with its several
thousand graves that unfortunately has never been published.9
However, none of the afore-mentioned cities yielded nearly as
many finds as Carthage.
When compared with other Late Antique cities of the
Mediterranean, Carthage holds a special position. An overview
of Byzantine buckles of the late 6th and 7th centuries reveals
that even comparatively well known sites such as Corinth or
Sardis yielded hardly more than 20 pieces respectively.10 One
‘Intelligible Beauty’ | 133
Eger
Plate 1 Buckle from Hippo Regius/Annaba, Algeria.
(scale 1:1)
Plates 2–3 Buckles of ‘Syracuse’ type, variants 1 and 2, Carthage (?) (scale 1:1)
must ask whether the finds in the Musée National de Carthage
all come from Carthage itself or from other ancient sites in
Tunisia and eastern Algeria. Although this problem cannot
be dealt with in detail here, it remains to be recorded that
the extraordinarily large collection of the Musée National de
Carthage and some dozens of pieces from other important
North African sites form a sufficient data base for the questions
addressed in the introduction. Consequently both the
potentials and limitations of regional differentiation will be
explained by selected examples of Byzantine buckles.
Byzantine buckles of the Vandal period
It is mainly due to the existence of burials with grave goods,
which are likely to be connected to the élite of Vandal and Alan
society, that some examples of Mediterranean goldsmith’s work
of the middle and the second half of the 5th century survive.
Amongst these finds are, for instance, two grave groups from
Hippo Regius (mod. Annaba), now in the British Museum.
Of these finds I would like to argue that an oval buckle from
female grave 1 is a clear example of an East Mediterranean
Byzantine type (Pl. 1). The contents of this grave from Hippo
Regius were left to the British Museum in 1865 and have only
recently been discussed in detail by Quast. They include a pair
of cloisonné disc-brooches and a gilt-bronze belt buckle with a
looped belt-plate.11 The latter piece consists of an oval loop of
circular section, an undecorated club-shaped tongue of semicircular section, and a kidney-shaped sheet metal plate which
displays on the front a lion hunt executed in fine punch work.
On the reverse the two hinge-loops turn around to the lower
metal plate, which survives intact apart from minor damage
on the rim and is of identical size to the upper metal plate. The
belt was fixed between the two sheet metal plates by means of
three laterally placed rivets with small hemispherical heads.
The piece belongs to a well-investigated type, called IV.1.B
by Kazanski, Bône-Csongrád by Quast, and B10 by SchulzeDörrlamm.12 The buckles always consist of an oval loop, an
undecorated regularly broadening club-shaped tongue, and
an oval or kidney-shaped belt-plate. There are only a few
differences with regard to this type such as variations in size
between the upper and lower plates of the belt-plate. Whenever
the reverse is illustrated, the lower plate differs from the
upper one by being narrower, trapezoidal or tongue-shaped.
There are only a few exceptions to this rule such as a fitting of
134 | ‘Intelligible Beauty’
unknown provenance made by hollow casting and now in the
collection of the Römisch-Germanishes Zentralmuseum in
Mainz,13 and the piece presented here with an upper and lower
plate of identical size. The punched or engraved ornamentation
is clearly more individual and includes a wide variety of
figurative motifs, as well as Greek inscriptions. Simple lion
representations are found on three specimens from SyracuseGrotticelli, Asia Minor, and another piece of unknown origin.14
They resemble each other to such a degree that they may well
have come from the same workshop. No parallel, however, to
the lion hunt on the Annaba buckle exists.
The origin of these buckles in the Byzantine Greekspeaking East is beyond doubt, not least because of the
epigraphic pieces.15 The distribution map supports this
with surprisingly clear evidence (Pl. 14). Thanks to the six
specimens in the Römisch-Germanisches Zentralmuseum
whose provenance is given as ‘Asia Minor’, one focus can be
located here and another on the lower Danube river, more
exactly in the Dobruja.16 Other examples have been noted in
Sicily. The buckle from Annaba is the most south-western findspot. North of the Alps, the type is completely absent, and on
the middle Danube only one piece is known, from Csongrád.17
The most important circumstantial evidence for a dating
exclusively in the 5th century is the presence of similarly
shaped undecorated buckles of precious metals from the
Merovingian Empire, which Wieczorek has dated to the
middle and last third of the 5th century.18 With its large lower
metal plate the buckle from Hippo Regius grave 1 displays a
typologically early characteristic, which places the piece at
the beginning of the life span of type Bône-Csongrád/B10.19
This fits well with the dating of the pair of disc-brooches to
the mid-5th century convincingly established most recently by
Quast.20
‘Byzantine’ buckles with fixed and hinged plates
By far the largest number of dress accessories from North
Africa belong to the Byzantine period and specifically to the
so-called Byzantine buckles with fixed or hinged plates. A
refined typology for buckles with fixed plates was established
by Schulze-Dörrlamm only a few years ago, and a similar
one for buckles with hinged plates is in preparation by her.21
Schulze-Dörrlamm collected numerous parallels in order to
define both the distribution and origin of the individual types.
Byzantine Dress Acessories in North Africa
Plate 4 Buckle of ‘Syracuse’ type, variant 3,
Carthage (?) (scale 1:1)
Plate 5 Buckle-plate from Carthage, Quartier Magon,
Tunisia (scale 1:1)
North African buckles, however, were largely unknown to her,
so several of her distribution maps require amendments in this
region.
Buckles of ‘Syracuse’ type
A type current all over the Byzantine Empire and far beyond
its borders was first introduced into the scholarly literature by
Werner under the type-name ‘Syracuse’ (Schulze-Dörrlamm’s
type D 12).22 The chronological framework of this buckle type
has been thoroughly discussed, most recently by Riemer and
Schulze-Dörrlamm. According to their analysis, these buckles
can mainly be dated to the late 6th to third quarter of the 7th
century.23 With more than 40 specimens, 31 of which come from
Carthage, these oval buckles with fixed sub-circular plates
with inverted semi-palmettes form the second largest buckle
group in North Africa.24 On the basis of the North African
buckles several easily distinguishable variants can be defined,
three of which are discussed below (Pls 2–4).25
Variant 1: On the loop, flanking the tongue-rest, punched
commas are sometimes found. The decoration of the plate
consists of two inverted semi-palmettes. Roundels with two
double leaves are situated between the loop and the plate (Pl.
2).26
Variant 2: Size and shape of the plate are identical to
variant 1. ‘Dot and comma’ ornament is consistently found
flanking the tongue-rest and on the roundels between the loop
and the plate. The plate decoration consists of two narrower
semi-palmettes in shallower relief; additional rudimentarily
worked semi-palmettes with comma ornament exist on the
internal rim side of the plate. A central ornament in the shape
of a stylised flower with lateral buds is found between the
semi-palmettes (Pl. 3).
Variant 3: The plate is of a broadly oval shape and there
are no roundels between the loop and the plate. The relief
decoration is closely related to variant 2. With only three
specimens in the Museum at Carthage, this is the most
infrequent of the three forms (Pl. 4).
Other buckles from Carthage are related to the ‘Syracuse’
type because of their similar profiles, but their decorative
patterns are completely different. Buckles of the ‘Syracuse’
type, can be found all over the Mediterranean and in many
areas on the northern periphery of the Byzantine Empire. The
distribution map contains more than 75 find-spots, some of
them cemeteries with several specimens (Pl. 15).27 The sites
Plate 6 Buckle-plate from Siurgus (?) (scale
1:1)
range from Sussex in southern England to the Caucasus and
Sasanid Iran in the east and upper Egypt in the south. There is
hardly any other type of Byzantine buckle which demonstrates
mass production so clearly as the Syracuse type, although
the picture is influenced by various factors: apart from the
practice of burial with grave goods, the state of excavation
and publication seriously obscures both the real distribution
patterns and parameters of use in Late Antiquity. Thus, the
great number of find-spots in the Crimea signifies both the
distinctive custom of burying with grave goods practised by
the population in the so-called Crimean-Gothic cemeteries,
as well as the frequent importation of buckles facilitated by
the immediate vicinity to Byzantine territory (here the city of
Cherson and its hinterland).
The distribution of the ‘Syracuse’ type within the Byzantine
Mediterranean must be appraised quite differently. Italy seems
to represent the ‘normal case’. Thanks to Riemer’s study of
Early Medieval graves in Italy, finds are documented more
or less evenly nationwide.28 Quite different is the situation in
the south-east and south of the Byzantine Empire, both areas
which appear almost totally to lack finds of this type. Here
the situation is aggravated by the lack of both studies of old
material in storerooms and of appropriate publications of small
finds from recent excavations.
With the new pieces from North Africa and also a few
specimens from Spain,29 the south-western Mediterranean is
now more clearly represented in the overall distribution of this
type. The great number of pieces from Carthage is particularly
notable and is likely to indicate local production. When the
distribution is differentiated according to variants, the first one
is by far the most frequent, also with regard to supra-regional
level. Its distribution pattern largely coincides with the overall
distribution. Variant 2 is clearly subordinate numerically, with
about a quarter of all finds, but also possesses a supra-regional
distribution hardly inferior to variant 1.
‘Syracuse’ buckles of variant 3 are rarer not only in
Carthage, but also in general. They have a clearly different
distribution, with all of the northern imperial territories and
the barbarian areas beyond devoid of finds of this type. With
the exception of two pieces from Italy, the few specimens
are exclusively limited to the east and south of the Empire.
Therefore, the workshops for variant 3 are likely to have been
situated somewhere in the wide crescent between Carthage,
the Levant and Asia Minor.
‘Intelligible Beauty’ | 135
Eger
Plate 7 Buckle from Siurgus (?) (scale 1:1)
Buckles of ‘Hippo Regius’ type
The second type of buckle are those with a hinged plate known
as the ‘Hippo Regius’ type. (Pl. 5). In this case the eponymous
finds were several buckles found by Marec during excavations
in the city area of Hippo Regius.30 This type was dated by
Riemer well into the 7th century, although it remains uncertain
whether the dating of this type extends into the last third of
the century.31 In Carthage these buckles form by far the largest
group. The 57 specimens are split between half with a larger
and half with a smaller variant of plate. Their characteristic
features are lateral convexities and a roundish end, which give
the fitting an almost horseshoe-shaped appearance.
The type belongs to those few buckles whose decoration
was produced not by casting, but by engraving and punching.
The iconographic repertoire includes zoomorphic and
anthropomorphic representations, more rarely monograms
or foliate decoration. Zoomorphic motifs are clearly dominant
with more than 40 examples. Amongst the most frequently
depicted animals are horses, birds and lions. Mythological
animals, such as griffins and Pegasus, are also found, but are
insignificant numerically. A uniform iconographic programme
cannot be identified. The spectrum ranges from merely
decorative images of individual animals, to secular themes
such as the world of the circus, to busts of saints and the pious
blessing formula of ‘light and life’ in monogrammatic form.
Their distribution (Pl. 16) spans almost the whole of
the Mediterranean and ranges from Spain to Asia Minor
and Egypt, where several specimens have only recently
been published.32 On the other hand, the ‘Hippo Regius’
type is unknown in the Crimea and the Black Sea region.33
Distribution patterns are concentrated, however, in Italy
and Northern Africa and therefore quite clearly indicate an
origin for this type in the central Mediterranean. Riemer
found examples in Sardinia, but also in Sicily and northern
and central Italy.34 Their frequent occurrence in Carthage, in
combination with a group of a dozen buckles from smaller
towns on the southern fringe of the province of Numidia,
such as Timgad, illustrates the particular importance of this
buckle type in North Africa. An attribution to North African
workshops can be assumed for this type, not only due to the
large numbers found there but also because of typological and
iconographic factors.
In North Africa, the images are all (with one exception)
orientated parallel to the longitudinal axis of the plate. When
136 | ‘Intelligible Beauty’
the buckles were worn these images were turned horizontally
so that an observer could not immediately recognize the
design. In Sardinia, on the contrary, several examples survive
on which the image was arranged at a right angle to the
longitudinal axis of the plate and would thus have been more
easily visible when worn. Further examples include a plate
with Daniel amongst the lions, from the province of Laerru, a
plate with two females between crosses (Pl. 6), and a buckle
with the figure of a blessing Christ from Siurgus (Pl. 7),35 all
unknown in a North African context, and which probably
represent the output of an independent workshop situated in
Sardinia. On the other hand, iconographic comparisons show
that individual North African motifs can also be paralleled
on other buckles from the Mediterranean region. Since the
decoration of the plate was engraved after the initial casting,
the identically executed pictures might hint at the work of one
and the same – in this case North African – engraver.
Buckles with a long plate with undulating profile and buckles
of ‘Riva San Vitale’ type
Apart from the numerous types referred to above, there are
also a number of unique pieces in the collection of the Carthage
Museum which are as yet unparalleled. This is the case with
a bronze buckle of long shape with undulating profile and
engraved with a biblical scene (Pl. 8) representing Abraham
and the three angels by the oak of Mamre, a story told in
Genesis 18:1–15. The iconography of this scene is well known
in the early Byzantine period as is demonstrated, for instance,
by a mosaic in the church of San Vitale in Ravenna. Two other
buckles share the same shape, but differ in their decoration.
One of them is also from Carthage and is decorated with Christ
and the 12 apostles, the other of an unknown Algerian origin
and now in the Musée National des Antiquités, Algiers, is
undecorated. All three buckles may be dated from the mid-7th
to the early 8th century due to the long narrow shape of their
plates. Exact counterparts from outside Northern Africa are
unknown.
Another unique piece is a gold buckle plate, broken into
two pieces, which represents the only gold Byzantine dress
accessory so far known from Carthage (Pl. 9).36 The buckle
was hollow cast and subsequently engraved and punched.
With regard to its shape and decoration, it belongs to a group
of buckles with hinged lyre-shaped plates which were first
collected by Werner and dated to the 7th century.37 Werner
used the term ‘lyre-shaped’ for two different types of plates:
on the one hand for figure-of-eight shaped belt-fittings with
an integral plate, often decorated with animals and termed
Plate 8 Buckle- plate from Carthage (?) (scale 1:1)
Byzantine Dress Acessories in North Africa
the ‘Trebizond’ type, and on the other hand for buckles with a
medallion-shaped terminal, which he discussed in connection
with a buckle from Riva San Vitale, Switzerland.38 Both forms
are loosely related to each other and share some decorative
elements such as a beaded rim, ‘dot and comma’ ornament,
and decorative bosses. However, the two forms need to be
differentiated because the designs of their front sides are
different. Those buckles of the ‘Riva San Vitale’ type have a
quadripartite field, but with two different terminals, one with
a round, the other with a tear-drop shaped element. With
regard to the last variant, I would draw particular attention to
those examples with a combination of geometric and foliate
decoration and a beaded profile, as exemplified by the beltfitting from Carthage.39 They include several gold buckles
from, inter alia, Italy (Pl. 10), Asia Minor, and from unknown
Mediterranean sites.40 They have all come from the antiquities
trade. None of the aforementioned pieces is exactly identical
with the belt-fitting from Carthage. In most cases the faceted
bosses, characteristic of the Carthage piece, are absent. The
best parallel for the Carthaginian example is, therefore, the
gilt-bronze buckle with a rounded terminal from Riva San
Vitale. Similar bosses, usually spiral-shaped rather than
faceted, are known from different Hispano-Roman derivatives
of this type (Pl. 11).41 But these decorative bosses cannot be
explained as purely western Mediterranean features. Although
absent from specimens of the ‘Riva San Vitale’ type in the
eastern Mediterranean, decorative bosses are not unknown
there, as indicated by a gold buckle of the ‘Trebizond’ type in
a treasure from Constantinople42 or by a large gold cruciform
buckle now in the British Museum.43 Thus, the evidence seems
random. Gold buckles were precious accessories, which were
surely produced in very small numbers, or as individual
productions. Due to their small number and widely scattered
find-spots, it remains uncertain where they were made. In the
literature, Constantinople is usually considered to be the origin
of high-quality goldsmith’s work. However, Roth assumed that
the gold buckle plate mentioned above was made in a workshop
in Carthage.44 This is certainly a possibility as Carthage, the
centre of the south-western Mediterranean and the seat of a
Byzantine exarch, certainly housed specialised metal workers
and also goldsmiths.
Local production of dress accessories in Carthage?
An atelier for polychrome artefacts probably existed by the 5th
century in Carthage.45 This is indicated by a hoard of partly
pre-cut almandine fragments in the Museum of Carthage.46
Additionally, there is some evidence in Byzantine times for
local fine metal working and goldwork in Carthage. Roth drew
attention to a bronze model for forming sheet metal pectoral
crosses found during the American excavations at Carthage
- Dermech, which can be dated to the late 6th to 7th century
because of its palmette decoration and ‘dot and comma’
ornament (Pl. 12).47 Furthermore, there are several casting
moulds in the Museum of Carthage which clearly served for
Plate 9 Buckle-plate from Carthage (?) (scale 1:1)
Plate 10 Gold buckle from unknown site, Italy (scale 1:1)
Plate 11 Buckle from unknown site, southern Spain
(scale 1:1)
‘Intelligible Beauty’ | 137
Eger
Plate 12 Bronze model, Carthage-Dermech, ‘ecclesiastical complex’ (scale 1:1)
Plate 13 Casting mould, Carthage, Quartier Magon, Tunisia (scale 1:1)
the fabrication of cross pendants and small medallions with
cruciform monograms. Two casting moulds found during
the German excavations in room 31 of the so-called ‘Quartier
Magon’, together with fragments of copper-alloy waste, also
hint at the activity of a metalworker’s workshop (Pl. 13).
Unfortunately, all particulars for this workshop evidence,
which is unique in Carthage so far, are lacking.48
Proof for the local production of buckles, on the other
hand, also exists. One of the very few relevant pieces is a small
hinged fitting of the ‘Bologna’ type: it survives as a cast blank,
the ridges of which have not been removed and hinge eyes and
perforated studs have not yet been drilled. The piece was nonfunctional and likely to have been found in a metal workshop.
Unfortunately, it belongs to the old finds in the Musée National
de Carthage so there is neither a find-spot nor context for it,
although a Carthaginian origin is not unlikely. As a fitting of
the ‘Bologna’ type, it belongs to one of the most ubiquitous
types of Byzantine buckle. It proves not only that Byzantine
buckle types with a distribution pattern indicative of regional
production were made locally, but also that this can be true for
types with a circum-Mediterranean distribution.
distribution pattern. Buckles with an oval loop fitting of the
‘Bône-Csongrád’ type have a mostly eastern Mediterranean
distribution, which only reached Sicily and sporadically North
Africa. The ‘Syracuse’ type was presented as an example of
a widely distributed type, present throughout most of the
Mediterranean and far away beyond the northern imperial
borders. A more detailed subdivision of the type shows,
however, that less widely distributed variants occurred. With
buckles of the ‘Hippo Regius’ type on the other hand, one sees
a characteristic central Mediterranean and particularly North
Africa type. The three buckles with elongated plates with
undulating edges, including the buckle with Abraham, may
be considered North African regional types. Because of the
present state of publication, it is not always possible to specify
regional attributions. This was illustrated also by the fourth
type (‘Riva San Vitale’), a lyre-shaped gold belt-fitting with
typical Byzantine elements of shape and decoration, which is
unique and has thus escaped a closer regional allocation. Still,
it cannot be excluded that it was made in North Africa, possibly
even in Carthage itself. Finds of casting moulds for different
pieces of jewellery and the cast blank of a buckle-fitting attest
to the activity of metalworkers in the North African metropolis.
The evidence provided by the examples discussed above
indicates that belt buckles from North Africa of the 5th to 7th
centuries include genuinely eastern Mediterranean, central
Mediterranean and probably also special North African types.
I have not been able to deal here with certain types present
both in the central and western Mediterranean or with (north-)
western Mediterranean and Pontic ‘imports’. For all regional
attributions one must keep in mind that we are not dealing
with sharp borders. This becomes particularly clear with
regard to buckles of the ‘Hippo Regius’ type: despite a clear
concentration of find-spots in North Africa and southern
Italy the overall distribution reaches very much further and
selectively spans almost the entire Byzantine Empire. This
makes manifold connections visible, which I would tend to
explain – on the basis of the present find-spot distribution –
more to personal mobility than either trade or the distribution
of fashions. Buckles of the ‘Syracuse’ type, on the other hand,
could be interpreted as an example of a ‘fad’, because they
are spread literally all over the Mediterranean and their
production was quickly adopted in many places, presumably
Summary
The products of the minor arts of Late Antiquity tend to be
termed ‘Mediterranean’ or ‘Byzantine’ by researchers in a
general way. Nevertheless, in order to be able to describe
the development of craftwork and the relationship between
different regions of the Mediterranean, a stronger regional
differentiation is desirable as the state of sources and
publications improves. For the investigation of such questions,
a sufficient data base of the relevant material is necessary.
Therefore, belt buckles are a particularly suitable subject for
research as many hundreds of examples of the 5th to 7th/8th
centuries are now known throughout the Mediterranean.
North Africa itself, however, has in the past yielded only a
few finds. By the study of Late Antique dress accessories from
Carthage and other North African museums this find gap has
now been filled. From this rich material five types have been
discussed above: one of the 5th century (‘Bône-Csongrád’
type) and four of the later 6th and 7th centuries (‘Syracuse’,
‘Hippo Regius’, ‘Riva San Vitale’ and buckles with a long fitting
with undulating edges), each of which represents a different
138 | ‘Intelligible Beauty’
Byzantine Dress Acessories in North Africa
by passing on models and casting moulds. Although there is no
space here to go any deeper into this matter, it becomes quite
clear that belt buckles are ideally suited for the study of the
mechanisms which lead to koiné and regionality in Byzantine
metalwork.49
Acknowledgements
13
14
I would like to thank Dr Janine Fries-Knoblach, Munich/Dachau, for
reading and translating the manuscript from the German.
15
Notes
Cf. J. Drauschke, ‘Zur Herkunft und Vermittlung ‘byzantinischer
Importe’ der Merowingerzeit in Nordwesteuropa’, in S. Brather
(ed.), Zwischen Spätantike und Frühmittelalter. Archäologie des 4.
bis 7. Jahrhunderts im Westen (Ergänzungsbände zum Reallexikon
der Germanischen Altertumskunde 57), Berlin/New York, 2008,
367–423.
2 Ch. Boube-Piccot, Les bronzes antiques du Maroc IV. L´équipment
militaire et l´armement, Paris, 1994; on Late Antique dress
accessories of the Tingitana cf. also, N. Villaverde Vega, Tingitana
en la antigüedad tardía (siglos III–VII). Autoctonía y romanidad en
el extremo occidente mediterraneo, Madrid, 2001.
3 But one should note the publication of Roman and Late Antique
fibulae by R.R. Gerharz, ‘Fibeln aus Afrika’, Saalburg Jahrbuch 43
(1987), 77–107, and the catalogue-like publication of graves and
isolated finds of the Vandal period by G.G. Koenig, ‘Wandalische
Grabfunde des 5. und 6. Jahrhunderts’, Madrider Mitteilungen 22
(1981), 299–360. Both authors, however, restricted themselves to
assembling published material only.
4 Koenig (n. 3); Ch. Eger, ‘Vandalische Grabfunde aus Karthago’,
Germania 79/1 (2001), 347–90; Ch. Eger, ‘Silbergeschirr und
goldene Gürtel. Die vandalische Oberschicht im Spiegel der
Schatz- und Grabfunde Nordafrikas’, Antike Welt 35/2 (2004),
71–6; J. Kleemann, ‘Quelques réflexions sur l’interprétation
ethnique des sépultures habillées considérées comme
vandales’, Antiquité Tardive 10 (2002), 123–9; D. Quast,
‘Völkerwanderungszeitliche Frauengräber aus Hippo Regius
(Annaba/Bône) in Algerien’, JbRGZM 52 [2005] (2007), 237–315;
Ph. von Rummel, Habitus Barbarus. Kleidung und Repräsentation
spätantiker eliten im 4. und 5. Jahrhundert (Ergänzungsbände
zum Reallexikon der Germanischen Altertumskunde 55), Berlin/
New York, 2007; Ch. Eger, ‘Vandalisches Trachtzubehör? Zu
Herkunft, Verbreitung und Kontext ausgewählter Fibeltypen
aus Nordafrika’, in G. Berndt and R. Steinacher (eds), Das Reich
der Vandalen und seine (Vor-) Geschichten (Forschungen zur
Geschichte des Mittelalters 13), Vienna, 2008, 183–95.
5 E. Marec, ‘Hippone: Objets en bronze récemment découverts’,
Libyca 6 (1958), 163–71.
6 Cf. M. Mackensen, ‘Metallkleinfunde’, in F. Rakob (ed.), Karthago
III. Die deutschen Ausgrabungen in Karthago, Mainz, 1999, 530–44.
Very few buckles were found in either of the British excavations:
H.R. Hurst and S.P. Roskams, Excavations at Carthage: The
British Mission vol. 1,1. The Avenue du President Habib Bourguiba,
Salammbo: the site and finds other than pottery, Sheffield, 1984;
H.R. Hurst, Excavations at Carthage: The British Mission vol.
2,1. The circular harbour, north side. The site and finds other than
pottery (with contributions by C. Duhig et. al.), Oxford, 1994.
7 Delattre published only three Byzantine buckles. Two of them,
however, he believed to have originated in the reign of St Louis
IX. This French king set off on a crusade to the Holy Land in 1270
and landed near Carthage where he died from the plague shortly
afterwards (cf. A.-L. Delattre, Souvenirs de la croisade de Saint
Louis trouvés à Carthage, Tunis, 1890).
8 Ch. Eger, Spätantikes Kleidungszubehör aus Nordafrika (in
preparation); Cf. provisionally, Ch. Eger, ‘Boucles de ceinture de la
région de Carthage datant des VIe et VIIe siècles’, CÉDAC Carthage
Bulletin 19 (1999), 12–15.
9 Briefly mentioned e.g. by, Ch.Courtois, Timgad. Antique
Thamugadi, Algiers, 1951, 66–7.
10 Only buckles of the late 6th and 7th century were quantified.
11 Quast (n. 4), 239–47.
12 M. Kazanski, ‘Les plaques-boucles méditerranéennes des Ve–VIe
siècles’, Archéologie Médiévale 24 (1994), 161–2: buckle with oval,
respectively kidney-shaped fitting and figurative decoration.
1
16
17
18
19
20
21
22
23
24
25
For buckles with epigraphic decoration: Quast (n. 4), 258–9,
fig. 16, 305, find-list 1; M. Schulze-Dörrlamm, Byzantinische
Gürtelschnallen und Gürtelbeschläge im Römisch-Germanischen
Zentralmuseum I. Die Schnallen ohne Beschläg, mit Laschenbeschläg
und mit festem Beschläg des 5. bis 7. Jahrhunderts (Kataloge vorund frühgeschichtlicher Altertümer 30), Mainz, 2002, 54–9.
Schulze-Dörrlamm (n. 12), 57, no. 51.
Koenig (n. 3), 345, fig. 18b; Schulze-Dörrlamm (n. 12), 56, no. 49; L.
Wamser (ed.), Die Welt von Byzanz– Europas östliches Erbe. Glanz,
Krisen und Fortleben einer tausendjährigen Kultur, Munich, 2004,
279, no. 437.
Quast (n. 4), 259, fig. 16 distinguished between figurative and
epigraphic decoration, but found no significantly different
distribution, particularly since another example with a Greek
inscription from eastern Sicily has been published in the meantime
(Manganaro Collection: cf. G. Manganaro, ‘Arredo personale del
Bizantino in Sicilia (fibbie, spille, anelli)’, Byzantino-Sicula IV. Atti
del I Congresso Internazionale di Archeologia della Sicilia Bizantina,
Palermo, 2002, 506, fig. 8,106).
Quast (n. 4), 259, fig. 16. These are to be complemented by the
following examples: unknown provenance, eastern Sicily
(Manganaro [n. 15], 506, fig. 8,106,108); Uluköy-Hocalar, Turkey
(M. Lightfoot, ‘Belt buckles from Amorium and in the Afyon
Archaeological Museum’, in C.S. Lightfoot (ed.), Amorium Reports
II. Research Papers and Technical Reports (BAR International Series
1170), Oxford, 2003, 99, pl. VI/22); unknown provenance, environs
of Mersin (?), Turkey (R. Taştekin, Früh- und mittelbyzantinische
Trachtbestandteile in der Türkei (Unpublished Master of ArtsThesis), Bonn, 1996, 81, pl. 11, fig. 1a–b); unknown provenance
(Wamser [n. 14], 279, no. 438); Chersonese, Crimea/Ukraine,
cistern P/1967 (I. Gavrituchin, ‘Fibuly i remennye garnitury
iz cisterny p-1967 g. v Xersonese‘, Materialy po arxerologii,
istorii i entnografii Tavrii 9 (2002), 225, fig. 1,8). With the last
mentioned example, Crimea has henceforth been included into the
distribution area.
D. Csallány, Archäologische Denkmäler der Gepiden im
Mitteldonaubecken (454–568 u. Z.), Budapest, 1961, pl. 210,11. Cf. D.
Quast, ‘Byzantinisch-gepidische Kontakte nach 454 im Spiegel der
Kleinfunde’, in E. Istvánovits and V. Kulcsár (eds), International
Connections of the Barbarians of the Carpathian Basin in the 1st–5th
centuries ad. (Proceedings of the International Conference held
in 1999 in Aszód and Nyiregyháza), Aszód and Nyiregyháza, 2001,
437.
A. Wieczorek, ‘Die frühmerowingischen Phasen des Gräberfeldes
von Rübenach. Mit einem Vorschlag zur chronologischen
Gliederung des Belegungsareales A’, Berichte der RömischGermanischen Kommission 68 (1987), 410–16. Cf. SchulzeDörrlamm (n. 12), 56–9. Cf. also a recent find from Chersonese
(Gavrituchin [n. 16], 225, fig. 1, 8) which is unlikely to be older than
the late 5th century.
Cf. however, J. Kleemann, ‘Rezension zu Schulze-Dörrlamm,
Byzantinische Gürtelschnallen’, Ethnographisch-Archäologische
Zeitschrift 46 (2005), 117, who takes the tiny shape of the topplate amongst other things as circumstantial evidence for an
independent western production.
Quast (n. 4), 248–50.
Schulze-Dörrlamm (n. 12), 2002. See: eadem, Byzantinische
Gürtelschnallen und Gürtelbeschläge im Römisch-Germanischen
Zentralmuseum II, Mainz, forthcoming, for her analysis of pieces
with hinged (moveable) plates.
J. Werner, ‘Byzantinische Gürtelschnallen des 6. und 7.
Jahrhunderts aus der Sammlung Diergardt’, Kölner Jahrbuch für
Vor- und Frühgeschichte 1 (1955), 36–48, at 37; Schulze-Dörrlamm
(n. 12), 171–9.
Schulze-Dörrlamm (n. 12); cf. also, Ch. Eger, ‘Eine byzantinische
Gürtelschnalle von der Krim im Hamburger Museum für
Archäologie’, Materialy po arxeologii, istorii i etnografii Tavrii 5
(1996), 343–7. The early 6th century date is given in E. Riemer,
‘Byzantinische Gürtelschnallen aus der Sammlung Diergardt
im Römisch-Germanischen Museum Köln’, Kölner Jahrbuch
für Vor- und Frühgeschichte 28 (1995), 777–809, at 779, and,
eadem, Romanische Grabfunde des 5.–8. Jahrhunderts in Italien
(Internationale Archäologie 57), Rahden/Westf, 2000, 149–50,
and is based on a single grave find from Sicily.
Only the ‘Hippo Regius’ type is more numerous, see below.
However, I have abandoned my previous typologies/variants
‘Intelligible Beauty’ | 139
Eger
26
27
28
29
30
31
32
33
34
35
36
as formulated in 1996 because this would imply a relative
chronological typo-genesis, which cannot be proven, namely
a dependence of variants 2 (termed variant 1 in 1996) and 3 on
variant 1.
Variant 1 is identical with the basic type according to Eger (n. 23).
Schulze-Dörrlamm 2002 (n. 12), 176, fig. 62.
Riemer 2000 (n. 23). Riemer’s study, however, was restricted to
published finds.
Ceuta: Villaverde Vega (n. 2), 497, Taf. 9, PT-81; Cartagena: La vida
en Bizancio en Carthago Spartaria. Aspectos de la vida cotidiana
(5 de abril – 31 de julio 2005. Museo Arqueológico de Cartagena
‘Enrique Escudero de Castro’), Cartagena, 2005, 98, no. 41.
Marec (n. 5); Riemer 1995 (n. 23), 790–1 was the first to describe
the type as ‘Hippo [Regius]’.
Cf. Riemer 1995 (n. 23), 791; Riemer 2000 (n. 23), 217–8.
Cf. Riemer 1995 (n. 23), 791. For examples from Egypt see: E. M.
Ruprechtsberger, ‘Eine byzantinische Gürtelschnalle aus Ägypten
mit Inschrift und figuraler Darstellung’, in U. Horak (ed.), Realia
Coptica. Festgabe zum 60. Geburtstag von H. Harrauer, Vienna,
2001, 95; J.L. Bacharach, Fustat Finds. Beads, Coins, Medical
Instruments, Textiles and Other Artifacts from the Awad Collection,
Cairo, 2002, 201, figs 2–3.
Why Ripoll (G. Ripoll, Toréutica de la Bética (siglos VI y VII d. C.),
Barcelona, 1998, 187) assumed a distribution also along the Black
Sea coast remains a mystery.
Riemer 2000 (n. 23), 217.
L. Pani Ermini and M. Marinone, Catalogo dei materiali
paleocristiani e altomedievali (Museo Archeologico Nazionale di
Cagliari), Rome, 1981, 114–15, nos 189–90. For a parallel for the
first example, see R. Temple (ed.), Early Christian and Byzantine
Art, London, 1990, 108, no. 77 (from Asia Minor).
W. Gernhöfer, ‘Die Darstellungen der drei Männer an der Eiche
von Mamre und ihre Bedeutung in der frühchristlichen Kunst,
140 | ‘Intelligible Beauty’
Römische Quaralschrift 104/1–2 (2009), 1–20.
37 Cf. H. Roth, ‘Almandinhandel und – verarbeitung im Bereich des
Mittelmeeres’, Allgemeine und vergleichende Archäologie Beiträge
2, 1980, 330–1 (where the material is wrongly described).
38 Werner (n. 22), 36.
39 J. Werner, ‘Byzantinische Gürtelschnalle aus Riva San Vitale (Kt.
Tessin)’, Sibrium 3 (1956/7), 79–84.
40 Individual pieces have a figurative decoration. These, and buckles
with a representation of Christ derived from this type, also remain
unconsidered here.
41 For the piece from Italy, cf. Riemer 2000 (n. 23), 219, fig. 25.
42 Cf. Werner (n. 39), pl. E; also Ripoll (n. 33), 157, fig. 27,113–4.
43 M.C. Ross, Catalogue of the Byzantine and Early Mediaeval
Antiquities in the Dumbarton Oaks Collection 2. Jewelry, enamels
and art of the Migration Period. With an addendum by S.A. Boyd
and St. R. Zwirn, Washington DC (2nd ed.), 2005, pl. 10, 4F.
44 Cf. Sotheby & Co., London. Sales catalogue 18 June 1968, pl.
facing 9; Schulze-Dörrlamm (n. 12), fig. 3. See also Entwistle, this
volume, no. 2.
45 Roth (n. 37), 331.
46 Ibid., 326–7, pls 5A–B.
47 Ibid., 330 fig. 7, 1; cf. W.H.C. Frend and J.H. Humphrey, ‘Bronze
Cross and Christian Inscription’, in J. H. Humphrey (ed.),
Excavations at Carthage 1976 conducted by the University of
Michigan III, Ann Arbor, 1977, 167–8, figs 1–2, who did not
recognize the actual function of the Carthaginian example.
48 With regard to a dating in the first half of the 6th century, cf. the
synoptical table by O. Teschauer in F. Rakob (ed.), Karthago I. Die
deutschen Ausgrabungen in Karthago, Mainz, 1991, 210. However, a
more detailed description and analysis of the Late Antique phases
of insula E117 are lacking.
49 Cf. also Drauschke (n. 1), 416–22.
Algeria
1. Hippo Regius/Annaba (grave 1/1865)
Bulgaria
2. Abritus/Razgard
3. Novae/Svištov
4. Novakovo
5. Unknown
Italy
6. Agro di Rugitgliano
7. Chiramonte-Gulfi
8. Syracuse-Grotticelli
9. Unknown, eastern Sicily
Romania
10. Callatis/Mangalia Gr. 8
11. Noviodunum/Isaccea
12. Piatra-Frecaţei Gr. D31
13. Tomis/Constanţa
Ukraine/Crimea
18. Kerč, Cistern P/1967
Turkey/Asia Minor
14. Sardis
15. Uluköy-Hocalar
16. Unknown, Asia Minor
17. Unknown, Mut near Mersin (?)
From unknown sites
20. Unknown, Schmidt Collection, Munich
21. Unknown, RGZM, Mainz
22. Unknown, Ashmolean Museum
Hungary
19. Csongrád
Plate 14 Distribution map of buckles of type ‘Bône-Csongrád’/B10 (double circle: 5 and more examples; brackets: unknown site; number of site).
Plate 15 Distribution map of buckles of type ‘Syracuse’/D12 (double symbols: 5 and more examples; brackets: unknown site; circle: variant 1; triangle: variant 1a; rhombic: variant 2; square: variant 3; diagonal cross: variant 4; cross: other pieces, figures
missing or unknown by the author; number of sites)
Variant 1
Egypt
1. Unknown
2. Unknown
Algeria
3. Rummel valley near Constantine
4. Djemila
5. Timgad
Bulgaria
6. Sofia
Germany
7. Freilassing-Salzburghofen (grave 68)
8. Langenlonsheim (grave 447)
9. Regensburg-St. Rupert (grave)
10. Sömmerda (grave 7)
Greece
11. Athens (Agora grave 10)
12. Athens (Kerameikos grave h)
13. Daskaleio
14. Edessa
15. Corinth
16. Corinth-Kraneion
17. Samos (Klostergut graves 2, 3, 4)
18. Samos (Eupalinos tunnel)
Iran
19. Unknown, Iran (?)
Italy
20. Mattinata-Stinco, Monte Sacro
21. Mondugno
22. Neapel
23. Unknown, near Naples
24. Nocera Umbra (grave 156)
25. Ostia, Basilica near Piana Bella (grave 21)
26. Piana degli Albanesi-S. Agata (grave 51)
27. Rutigliano (grave 6/1983-84)
28. Syracuse (?)
29. Tharros
30. Unknown
Jordan
31. Gerasa
32. Unknown, region of Shobeq (?)
Croatia
33. Pula
34. Salona
35. Zadar
Libya
36. Sabratha
Poland
37. Unknown, northern Poland (?)
Romania
38. Balta Verde
Russia
39. Boriszovo-Djurso near Novorosijsk
Spain
40. Cartagena (theatre)
41. Ceuta
42. Unknown, southern Spain (?)
Tunisia
43. Aïn Wassel
44. Unknown, Carthage (?)
45. Unknown, central Tunisia
Turkey
46. Amorium
47. Anemurium
48. Ephesus (Hanghaus 2)
49. Istanbul
50. Nysa
51. Pergamon
52. Sardis
53. Unknown
54. Unknown, Asia Minor
Ukraine, Crimea
55. Aromat (graves 1 and 6)
56. Bakla (grave 320)
57. Cherson
58. Čufut Kale (grave 7)
59. Eski Kermen (grave 257)
60. Feodosia
61. Kerč
62. Skalistoe (graves 258, 279, 320, 471)
63. Suuk Su (graves 58, 157)
64. Usen Bash (grave 2)
Iran
80. Unknown, Iran (?)
Tunisia
103. Unknown, Carthage (?)
Italy
81. Agrigent near San Leone
82. Calvisano-Santi di Sopra (grave Q61)
83. Sofiana (?), Sicily
84. Unknown
Turkey
104. Istanbul
105. Unknown, Asia Minor
Hungary
65. Csákberény-Orondpuszta (grave 215)
66. Kölked Feketekapu A (grave 325)
67. Szelevény
68. Szeged-Fehértó A (grave 34)
69. Unknown
Lebanon
86. Byblos or Tyros
Cyprus
70. Salamis
Of unknown origin
71. Northern Black Sea
LocaL imitation of Variant 1
Germany
72. Aubing (grave 167)
Italy
73. Aldeno
Variant 4
Croatia
85. Novigrad
Algeria
97. Rummel valley near Constantine
Italy
98. Luni
Libya
87. Sabratha
Uncertain
Tunisia
88. Unknown, Carthage (?)
Turkey
89. Amorium
90. Mut near Mersin?
91. Pergamon
Ukraine, Crimea
92. Aromat (graves 1, 6)
93. Eski Kermen (grave 218)
94. Kerč
95. Sacharnaja Golovka (grave 5)
Hungary
96. Kiszombor
Belgium
106. Belvaux
Germany
107. Stockstadt
Greece
108. Unknown
Italy
109. Brianza
110. Cividale
111. Perugia
112. Near Ravenna
113. Rome
114. Voltago
115. Unknown, Sardinia
Variant 2
(Nos 97–98 see below)
Germany
74. Hahnheim
Variant 3
Greece
75. Plateia, Argolis
Egypt
99. Antinoë
Great Britain
76. Chichester the Broyle
77. Colchester
78. Unknown, Kent
79. Unknown, Sussex
Israel
100. Beth Guvrin
Ukraine, Crimea
118. Gursuff
119. Lučistoe
120. Suuk Su (Graves 29, 32)
Italy
101. Venice, lagoon
102. Unknown
Hungary
121. Dunapentele
122. Százhalombatta
Russia
116. Nalčik
117. Taman
Plate 16 Distribution map of buckles of type ‘Hippo Regius’ (double circle: 5 and more examples; brackets: unknown site; number of site)
Egypt
1. Cairo-Fustat
2. Unknown
Algeria
3. Annaba
4. Djemila
5. Khemissa
6. Matifou
7. Tiddis
8. Timgad
9. Unknown, ‘Numidia’
10. Unknown, Rummel valley near Constantine
11. Unknown, eastern Algeria (?)
Germany
12. Prien am Chiemsee
13. Weißenburg
France
14. Near Belfort
Greece
15. Unknown
Israel/Palestine
16. Unknown
Italy
17. Civezzano
18. Faenza (grave 2)
19. Luni
20. Perugia
21. Rome-Crypta Balbi
22. Unknown, Etruria
23. Unknown, Lavello (?)
24. Unknown
25. Unknown
Italy, Sardinia and Sicily
26. Canicattini
27. Laerru
28. Noto
29. San Marco d’Alunzio
30. Siurgus (?)
31. Siurgus Donigala-Su Nuraxi
32. Is Pirixeddus near Sant’Antioco
33. Sulcis near Sant’Antioco (?)
34. Unknown, Sardinia
35. Unknown, Sicily
36. Unknown, eastern Sicily
Spain
37. Alayor, Menorca
38. Castell de Santueri, Mallorca
39. Italica
40. Torre Llisa Nou, Menorca
41. Unknown, ‘Baetica’
42. Unknown
43. Unknown, Andalusia (?)
Tunisia
44. Carthage-Quartier Magon
45. Near Carthage-Byrsa
46. Carthage-near Kothon
47. Unknown, Carthage (?)
Turkey, Asia Minor
48. Istanbul
49. Unknown, Asia Minor
50. Unknown, western Anatolia
Hungary
51. Pecs
From unknown sites
52. Unknown, Diergardt-Collection
53. Unknown
54. Unknown, Cabinet des Médailles, Paris