-,

---~

East Central and Eastern

Europe in the Middle Ages,
450-1450
A:
General Editor

Florin Curta

VOLUME2

-----.. - - - - - - - -

~.

_____ ._____

___,, ____.____..

_~,.,.._vz----

The Other Europe

s,

in the Middle Ages

Avars, Bulgars, Khazars, and Cumans

Edited by

Florin Curta
with the assistance of Roman Kovalev

~ l-EG/[)<"

:; ~~"" ..,,..
.,.
?",
c.,

..

""

/6 8 ~

BRILL

LEIDEN. BOSTON
2008

,t

Cover illustration: A mask,on,horse pendant from Vratsa.

Courtesy of the History Museum in Vratsa.

Ihis book is printed on acid,free paper.

A CLF. record for trus book is available from the Library of Congress.

ISSN 1872,8103

ISBN 978 90 04 16389 8

Copyright 2008 by Koninklijke Brill NV, Leiden, Ihe Netherlands.

Koninklijke Brill NV incorporates the imprints Brill, Hotei Publishing,

IDC Publishers, Martinus NijhoffPublishers and VSP.

All rights reserved. No part of this publication may be reproduced, translated, stored

in a retrieval system, or transmitted in any form or by any me ans, electronic, mechani

cal, photocopying, recording or otherwise, without prior written permission from the

publisher.

Authorization to photocopy items for internal or personal use is granted by

Koninklijke Brill NV provided that the appropriate fees are paid directly to

lhe Copyright Clearance Center, 222 Rosewood Drive, Suite 910,

Danvers, MA 01923, USA.

Fees are subject to change.

PRINTED IN THE NETHERLANDS

__

>

,~ _ _ _ _

_ _ _ _ ._ _. _ _

.~M_"'. _ _ _ ~~

_.

~ .

_,,, _ _ _ _ _ _ ./!_~

I
I

"
11

'I

CONTENTS
Preface ..................................................................................................
Introduction
Florin Curta

ix
1

Conflict and coexistence: the local population of the Carpathian

Basin under Avar rule (sixth to seventh century) ..........................
Tivadar Vida

13

Avar chronology revisited, and the question of ethnicity in the

Ayar qaganate ......................................................................................
Peter Stadler

47

Sew remarks on the flow of Byzal1tine coins in Avaria and

Walachia during the second half of the sevel1th century ..............
Peter Somogyi

83

Bulgars in the Lower Danube region. A survey of the

archaeological evidence and of the state of current research
Uwe Fiedler

151

Avar-age metalworking technologies in the Carpathian Basin

(sixth to eighth century) ....................................................................
Orsolya Heinrich -Tamaska

237

Two worlds, one hoard: what do metal finds from the forest
steppe belt speak about? ... ..... ......................... ........... ........................
Bartlomiej Szymon Szmoniewski

263

The earliest Avar-age stirrups, or the "stirrup controversy"

revisited ... ..... ....... ................... .......... ................ .............. ....... ..... .........
Florin Curta

297

A note on the "Hungarian sabers" of medieval Bulgaria

Valeri Iotov

327

viii

CONTENTS

Danube Bulgaria and Khazaria as part of the Byzantine

oikoumene ............................................................................................
Veselina Vachkova

339

From 'steppe' to Christian empire and back: Bulgaria between

800 and 1100 .......................................................................................
Tsvetelin Stepanov

363

A broken mirror: the Kip\ak world in the thirteenth century

Dimitri Korobeinikov

379

The Cuman bishopric-genesis and evolution

Victor Spinei

413

References ...........................................................................................

457

List of contributors .............................................................................

483

Index ....................................................................................................

485

----~~----'-,-'--

-~_._-~~,._-------_.- ... ~---~

tery B in

DlIection.

-earrings
gravc 8);
:h (Kesz
:hristian
rpalota

AVAR CHRONOLOGY REVISITED, AND THE QUESTION OF

ETHNICITY IN THE AVAR QAGANATEl
Peter Stadler

~rybythe

lctery by
Irt,grave
Drawing
~, Mller

The Avar age (ca. 570 to ca. 800) was aperiod of great significance for
the early medieval history of Europe. The Avar qaganate was the cre
ation of an elite of nomadic horsemen of eastern origin. Its early history
is known from literary sources, but for the later part (ca. 700 to ca. 800),
very few, if any such sources are known. However, the Avar age can now
be studied in great detail on the basis of archaeological excavations of
cemeteries and, lately, of settlements as weIl. During the last fifty years or
so, considerably energy has been invested in sorting out a firm chronol
ogy for the archaeological assemblages of the Avar age. Even though the
chronology of Avar history seemed clearly anchored to known moments
in history, in fact only the date for the Avar conquest of the Carpath
ian Basin (568) has received general acceptance. By contrast, the end
of the Avar qaganate, an event historians place in the early 800s, has
been dated by various archaeologists at various points in time between
800 and 900. More often than not, such differences in understanding
basic chronology stern from conflicting views on the medieval history
of the region, themselves based on differing views of national(ist) histo
ries. For example, most prominent among scholars inclined to date the
end of the Avar qaganate as late as possible within the ninth century are
Hungarian archaeologists and historians who insist that the first genera
tion of Magyars in Hungary coexisted with the last generation of Avars.
While absolute dates for the chronology of the Avar age remain under
discussion, great progress has been achieved in establishing a relative
chronology of archaeological ass em blages, especially for the later parts
of the Avar age for wh ich no coin-dated assemblages have so far been
found. More than forty years ago, Ilona Kovrig, the grande dame of
Avar archaeology, has proposed a chronological model based on the
division of the Avar age into Early, Middle, and Late periods. 2 Her chro
nology has meanwhile been greatly improved with the assistance of an

1
2

An expanded version of this paper appeared in Stadler 2005.

Kovrig 1963.

48

PETER STADLER

ever-increasing number of new assemblages and computer-assisted

methods to order them chronologiea11y. The division into Early, Middle
and Late Avar periods has been accepted by virtua11y a11 scholars with an
interest in the Avar age, even though they tend to favor quite different
absolute dates for the beginning and end of each one of these periods.

The relative and absolute chronology of the Avar age

The refinement of Kovrig's chronology has been made possible by the
application of new methods, especially the development and improve
ment of the image database "Montelius:' Named after the Swedish
archaeologist Oskar Montelius (1843-1921),3 the database came into
being in Vienna in 1999 and already has over 500,000 images pertain
ing to prehistoric and early medieval assemblages in Europe, all entered
by some 60 archaeologists, students, and volunteers. The coverage is
almost complete for the Avar period, with over 140,000 published arti
facts. The database consists of a co11ection of images of archaeological
artifacts allowing for the display of data in at least two different modes.
On one hand, the complex-view mode is not very different from the way
in whieh new archaeological information is presented visually in most
publieations, namely ordered by means of closed -find units (burial, set
tlement feature, or hoard assemblages). Figure 1 shows just one such
example, a Browser ACD. See image displaying artifacts found in the
rieh Avar-age burial in Kunbabony, wh ich some regard as the tomb of
one of the seventh-century Avar qagans. 4 By contrast, in the typological
mode, artifact images are grouped by formal similarity, the basic proce
dure for working with typology. In the typological mode, image could
be manipulated with the Drag 'n Drop tool activated by the computer
mouse. All changes operated in the typology structure are immediately
brought to the "background" database. Figure 2 shows an example of a
typology-mode view of pots with a S-shaped comb-punch decoration.
A number of different functions provide support for the work on such
an enormous typology. To input the image of any one artifact into the

3 Oskar Montelius refined the concept of closed find first introduced by Christian
Jrgensen Thomsen and in the process laid thc foundations of typology as a key method
for archaeological research. See Montelius 1903.
4 T6th and Horvath 1992.

49

AVAR CHRONOLOGY REVISITED

.er-assisted
rl)', Middle
ars with an
:e different
, periods.

Q)

-s
Ei

o
.-t1

....'"u

ible by the
I improve
e Swedish
came into
es pertain
all entered
:overage is
~shed arti
laeological
~nt mo des.
,mtheway
~y in most
burial, set
tone such
:md in the
le tomb of
:ypological
!Sie proce
lage could
computer
lillediately
ample of a
lecoration.
rk on such
ct into the

t:

<':!

"0

2U

Q)

'"

'S:

>

~~

>< ::l

Q)::r:

~~

E
o

\:i
0
U.o

Q)

'<':!

-S..g
~

-.S
ECil
('ij

.~

>< ;::

Q).o
\:i
<':!

0;; e.o
.2 ~

" Q)
'"<':!
..g
"iii
Cl

<':!

E
~

51

b. Christian
lkeymethod

Q)

"0 ....

e
'gJ

50

PETER STADLER

~
..-r...""'Ci

?-t..:...;"';~

"k"";~l

.'-~

Figure 2. Image Database "Montelius': an example of the typological mode

view: ceramic ware with S-shaped, prick-like comb punch ornament.

database takes less than 60 seconds. In the typology mode, the search for
formal analogies for any artifact takes only about 30 seconds. The allo
cation of one image to an existing type takes a few seconds more. A new
type is created easily by creating a new folder. An existing type can easily
be split up into two or more sub-types. In conclusion, a great advantage
over conventional typological methods is that comparisons may thus be
made 100 times faster than normally.
Figures 3 and 4 display in a schematic way how images are entered
into the Image Database "Montelius" and evaluations obtained on that
basis. Figure 3 starts from the "raw" publications, either monographs
or articles. Illustration plates displaying assemblages are scanned, and
individual artifacts are then separated by means of image processing.
Every single artifact image is then described in the mask of the pro
gram MonteliusEntry. On the other hand, the "raw" publication is also
the source of written information, wh ich can be catalogued along with
artifact images. By means of the Montelius seetion of the program pack
age known as WinSerion, images can then be presented either in the
complex mode or the typological mode. Figure 4 shows what can be
expected from WinSerion, once the data is entered into the database.
WinSerion allows for various kinds of seriation, in order to reveal pat
terns in the archaeological material considered. Moreover, local or
global maps generated by means of AutoCad offer the opportunity of
mapping finds by means of a WinSerion embedded Geographical Infor
mation System feature. Furthermore, WinSerion enables the user to

Figure

Foaoi
;c:~

-.

7~1II
~

Find units
In Image

Browser

Publications
; Monographie<

Artkies

Browser

umode
unent.

weh for
Ihe allo
e.Anew
:an easily
d.vantage
ythus be
~

entered
ion that
lOgraphs
ned, and
ocessing.
the pro
m is also
ong with
lID pack
er in the
lt can be
database.
~veal pat
local or
tunity of
cal Infor
e user to

51

AVAR CHRONOLOGY REVISITED

Access

Figure 3. A model for the creation of the Image Database "Montelius" on the
basis of the published archaeological record.

Find units

in Image Browser

e
I
---Dmagb
ata aSe ----

Montelius

..,...

V~nsenoy

1opoSeriation

Map ofEumpe
etc.

(GIS)

-h.t::;;;;;:::====?1

age.....
Typesm
. Im"'--Excavation maps _

Analysis of the
N Nexl /"
Neighbours

L,L--\-----v

I~

Absolute
Chronology

l
I

'

Chronology

.).L.._ _ _ _-"

{:.=====::zJL_L~=-=--====;:!:::----

J I

...--- ~ Mapping

/ ...--- !-- AutoCad- / '

Browser

p r ;7==:t====?1

--.
- ;I

~.-/'

Relative
,Chrooology

110.

Condensation /

I /
V /

Seriation
'Petrification
'Redprocal
Averaging
Triangulation

,c.=======?1
"Wiggle Matching"

Absolute Data

I
Figure 4. A model of the possible uses of the Image Database "Montelius"
for archaeological studies.

~:;:~

/:.;0

52

PETER STADLER

evaluate and compare automatically all maps produced by such means.

This is done by means of an algorithm known as analysis of the N -next
neighbors" (ANN). Seriation produces relative chronologies, which can
then be compared with the results of the spatial analysis performed by
means of ANN. Absolute data allow a linkage between relative chronol
ogies and absolute dates. The methods applied (sequencing and wiggle
matching) are based on Bayesian statistics, and their primary purpose is
to turn a relative into an absolute chronology.
Over 61,000 burial assemblages are known so far for the entire Avar
age. However, seriation by reciprocal averaging is only possible for
slightly more than 4,000 male burial assemblages with some 3,600 arti
fact categories. Figures 5 and 6 display the results of that seriation. While
in Figure 5 columns represent types and rows are assemblages, the x and
y axes in Figure 6 show the eigenvectors of assemblages and artifact cat
egories. In each one of the two graphs, there is a point for every artifact
found in an archaeological assemblage, for a total of over 20,000 points.
For both figures, the beginning of the chronological sequence is set on
the upper left corner ofthe graph, with the end in the lower right corner.
Another method of seriation, which is similar in principle, but pro
duces visibly different types ofgraphs, is correspondence analysis (CA).
The results of seriations by correspondence analysis are displayed in Fig
ure 7 for all burial assemblages with male skeletons, and in Figure 8 for
those with female skeletons. Every triangle in these graphs indicates an
assemblage (grave). The bigger the triangle, the more different types are
to be found in that assemblage. A standard CA seriation should produce
a parabola-shaped distribution of triangles. This is dearly the case for
the seriation of male, but not so for the female graves. In fact, the seria
tion of female graves produces a pattern consisting of two parabola-like
distributions joined at the center of the graph. I shall return shortly to
the interpretation of this analysis.
The different methods produce similar results for the highly refined
relative chronology. I have previously attempted to calibrate the relative

5 The method was invented and developed by Jean Paul Benzecri and his team of
the laboratory for Mathematical Statistics at the University of Paris VI (Benzecri 1973).
See B0lviken et al. 1982; Shennan 1990,283-86. Following the publication in English
ofthe first book based on Benzecri's ideas (Greenacre 1984), the CA gradually made
its appearance in Scandinavian and British, later in American, German, and Austrian
archaeology. For exemplary applications to medieval archaeology, see Hines, Nielsen,
and Siegmund 1999.

Bq

!,;.;:, -1

~t[

::+:tl

~511

".

....

S3

AVAR CHRONOLOGY REVISITED

:h means.
he N-next
which can
urmed by
~ chronol
nd wiggle
purpose is

Beginn
100

Typen
400

600

SUO 1000 1200 1400 1600 1800 2000 2200 2400 2600 2800 3000 3200 3401.1 3600

200

1 0 0 1 .' ,+---~-+---~i------+----t-----~---+---+--"L

400

wo

ntire Avar
)ssible for
3,600 arti
ion. While
, thexand
rtifact cat
:ry artifact
100 points.
eis set on
5ht corner.
~, but pro
ysis (CA).5
yedin Fig
igure 8 for
Idicates an
lt types are
ldproduce
le case for
the seria
Clbola-like
shortly to

tly refined
ne relative

200+-~-c~t~~~~~~~~j+~--~------1c-----~--~~-~~~--t----~~

000

1000
300 ~t-~-~-~-~4~--~-~-~-:;:';"

~--~t-~~-~~~---t---~'~'~~+~~~~L 1200

1400

400

+--+.~.~._~~..:-:-"'-:

:L.

',1

1600

2
~
~

1300

soo

-+~-~-~-+~'----t~~~~,-"~,~+.~~~t-~--

--r---+---~-

2000

]
j:

2200

600

-+--

,+~~+--~-.F---'--+---;---+----:"- 2400

J~

2600

700

+-----+-----+---'-T---+-~-~__c~_':l~'c"~'--~+__-__+__--.J-

2800

3000

800

3200

3400

+-~-"'_+~"'-.'

-l----f

:,1:. ",-.-e.+- 3600

3800

bis team of
zecri 1973).
in English
lually made
Id Austrian
es, Nielsen,

~-L.-L..f--'---.L-.L..-i---L------..L---l+~LJ~-+_- . .LJ..-~+-~L..:--I--_+--i..--L-L-1-~~-

300

400

500

600

700

800

900

Sequenzdaten [Typen]

Figure 5. Seriation by reciprocal averaging of over 4,000

Avar-age male burials.

' ....

4000

tOOO

Ende

54

PETER STADLER
FIlkIOI

.....
+3

-2.0

-1.5

-1.0

::,:,0.5

O.O

+1.0

+1.5

-2.0

+2.5

+3.0

2.0

+2

+1

-1.5
rtl

-1.0
-1

-2

-0.5
-3
0.0

Ci'
500

.:""

600

-2

"

Q..

6:
k

0.5 ~

"1lr:r

-4

Ci'

"

700

<il

'""

1.0
+5

1.5

2.0

J
I

+4

+3

.2
.1

2.5

-1

-2

'"
~ ;;:

""

:;;
"

""

3.0

Figure 6. Seriation by reciprocal averaging of the eigenvectors of over 4,000

Avar-age male burials.

-)

-4

-5
-6
-6

-;

Figure 8..

55

AVAR CHRONOLOGY REVISITED

Vaktof ;; ,",uj ~'Achw Q"'9~11 F"Klor

KOfuFpoondemanlll,.s..>.Fu~

3650- T..UOll(

+4

+3
+J.O

-2.0

+2

+1

-15
.t{)

-1.0

-1

..

-2

-0.5
-3
0.0

"~

-4

0.5

..

-1

"

-4
+2

Tl

s{)

Figure 7. Seriation by correspondence analysis of over 4,000

Avar-age male burials.

iS

"c:
1;l;

Fe!:t,,! i~tf)(-AdlW"qeqetl~'alrn>lla\,lI)'-~Me

t'l

!\OIIt1fpondOllJlz,an"lyse'l"Ilhd(> 1517

1.0

rl'PfM 7,21

-3

-2

-1

-)

-1

-1

+S
+4

1.5
+3"-'-'~-~~'--"-----"-'-"

+2

2.0
+1

.tO

2.5
-1

-2

3.0
-)

over4,000
-5 -:-.-

-6
--6

-,

..,;
-4

{)

+'

+2

+)

+4

+5

Figure 8. Seriation by correspondence analysis of Avar-age female burials.

+6

~
:~

56

PETER STADLER

chronology thus obtained to an absolute chronology by means of con

temporary Byzantine gold coins, which have been found in about thirty
burial assemblages. However, the number of coins is too small, espe
cially when compared to the large number of assemblages considered
for analysis and will not increase without excavation of other several
thousands of new graves. As a consequence, it is statistically impossible
to obtain an accurate absolute chronology on the basis of coins alone.
I therefore moved onto more precise methods of independent dating,
namely radiocarbon. Unlike coins, the number of radiocarbon sampies
can be easily multiplied from already excavated graves. We collected
about 100 sampies from archaeologically well-dated burial assemblages
from Hungary and Austria. All radiocarbon measurements were done at
the Vienna AMS facility VERA. 6 We began by dating the collagen from
human bones. Collagen is stored in the skeleton only until about the
twenty-fifth year oflife, after which it can be reconstructed only from
deconstruction products of old collagen, that is without using any new
or "fresh" carbon. 7 The choice of sampies took into consideration the
possibility of checking radiocarbon dates against the evidence of coins
from the same assem blages that have been tested. The results were over
whelmingly the same, given of course the margin of error for standard
radiocarbon measurements. 8 Figure 9 and 10 illustrate the degree of over
lap between seriation and radiocarbon dating of 38 sampies from male
graves. The method used for comparison is known as "wiggle matching"
and was performed with Oxcal3.9 in a somewhat modified way.9 As the
radiocarbon method does not give good results for the eighth and ninth
centuries, sampies were only taken from burial assemblages roughly
dated between 568 and 700. Confronting seriation with radiocar
bon dates leads to the conclusion that the Late Avar period begins
in ca. 680, and not as previously assumed (mainly on the basis
of seriation) in 700 or 720. This results in a considerable shift to earlier

6 VERA is an acronym for Vienna Environmental Research Accelerator. The direc

tor of this facility is Walter Kutschera, whose name is well-known among USA-based
scholars involved in radiocarbon measurements. All sampIes have been prepared for
measurement by Eva-Maria Wild and her team.
7 Wild et al. 2000.
8 For the principles and problems of radiocarbon dating, see Taylor 1987. For its
impact on prehistoric archaeology, see Renfrew 1973.
9 See Bronk Ramsey 2001.

:~
,.
.

57

AVAR CHRONOLOGY REVISITED

ms of con

bout thirty
naH, espe
:onsidered
ler several
impossible
ns alone.
ent dating,
m sampies
~ collected
isemblages
ere doneat
lagen from
aboutthe
only from
19 any new
~ration the
ce of coins
were overIr standard
reeofover
from male
matching"
ray.9 As the
land ninth
es roughly
radiocar
iod begins
the basis
i to earlier

The direc-

I CSA-based

prepared fr

f'.

1987. Fr its

V_Sequence Avar redud poor agreement Abs.Chronology.38 smpb

I

,Firsl]i""

i>::E~/:N
IG;];

73.3%

"'!<.fliE'

ib;~i;11
ib~:j:il
I:TJU\}24

IG~p23t2

1
118.4 % 1

IX. "SO fmn

,1' 'ill ,oftlfl{)11l8.1%

:~~:1;il

'1'

ib~:i:~O

,nnOll

W6~9%

,rh

"~ smpls
.....

\'l~600%)}

rv~

liiIa.

..,!,;;,A.

:b;~i14

ATSOT"0195

ib;~;r

ATLeofo152

!K;:~:r

A pm!,! 0"1110.6%

~I"
~/.~
~-'\.A-

.4 TMoe.OO93

ATSm, !m.OO61

70.3%

!~;:1:[61

tlV

1",.u,,>o

1!6.0%

134.3%

.~~\.

;~\

'~''''. /',

.J~V,-j',

/;.l
;:&,"'-.J\

....t.: ',,'

1~~;1:4'5

~!;1;r

A TMo f.OO29

11

b;;1:r

ATY;6 n ,,: 79.9%

11

b~~;l6

A1'Mvf0079

84.3%

I:

6~Rfi:1J

AT.7illi"oI0J'

125.1%

~\

/.1'--./",

122.2%

/..6: , .. '\

..~'~>"\.

b~:i;r

ATMoto222

132.9%

/~J'-,J',

I1

~~:1;jO

A1'Mof0489

142.0%

J:A,V'-. f

li

VERA 305

A1'.M0f'0334

VERA 291

o4TMoiOO35

v'ERA 277

ATLeof.OO78

VERA 309

o41'.Moe.0471

I ./:J...~A,
.

I
I

./.&../'......1\,

,'1vc,/'-,
/'-.\,

200AD

136.9%

I 79.1%

;~./"

129.2%

/24.5%

Ib~:1;ll

"'

~.-./\

81.0%

;K;:1if1
iK;;1;F

A1'Moto268

~,

T 7i11!,! nne7 46.4%

'0'

J':.oI!.

91.7%

I1
11

BCiAD

~-

,0011

:b;~:r

1I

,96,9%

11

933 96.,

109.~.

IYE~(;I03 \ T ,Co
IG;;';
:13
!~~~:l~ A,

11

I
I
I
I

~:;f71~

, re~uced poor agreeme~t '

'V_Sequence

400AD

600AD

800AD

!{){)()AD

Calendar date

Figure 9. Wiggle rnatching of radiocarbon dates with sequence dates frorn the
seriation of Avar-age burial assernblages.

58

PETER STADLER

T~

V_Sequence Avar reduced poor agreement Abs.Chronology.38 smpls

V_Sequence Avar reduced poor agreement Abs.Chronology.38 smpls

IA~94.4%(A'c~60.0%)J

---1

Phase I

I
200AD

300AD

400AD

500AD

600AD

700AD

800AD

900AD

1000.'\D

1100AD

Calend.r date

Figure 10. Wiggle matching of radiocarbon dates with sequence dates from

the seriation of Avar-age burial assemblages.

dates for the previously accepted chronology of the Avar age. The shift

is certainly to be explained by the fact that the coins, on which previous

dates were based, were in circulation at the time of burial, while most

artifacts found in burial assemblages may have been manufactured and

acquired between the twentieth and thirtieth lifetime year of the person

with whom they were buried. In other words, the date of the burial is

later by a few years than the date of production and acquisition that

can be established for the artifacts. On a more general level, the shift to

earlier dates of the later segment of the Avar chronology undermines

all assumptions among Hungarian archaeologists ab out the coexistence

of the last Avars and the first generation of Magyars in the Carpathian

Basin. In the light of the revised chronology, the end of ''Avaria'' must

now be placed shorty after 800, perhaps as late as 822, even though no

direct dates are available so far. By the same token, the beginning of the

Middle Avar Period is set at ca. 630.

Table 1 displays the overall effects of the new revised chronology of

Avar burial assemblages obtained by means of combining seriation with

radiocarbon dates. The column "Years AD 1" shows the new dates in

contrast to the old chronology displayed in column "Years AD 2:'

and

59

AVAR CHRONOLOGY REVISITED

Table 1: The chronology of the Avar age according to a combination of

seriation and radiocarbon dates
Years
ADl

Years
AD2

EAI

568

600

90

Early Avar 1I

EAII

Middle Avar I

MAI

600
630

630
655

90
180

180
360

Middle Avar II

MAI!

655

680

360

550

Late Avar I

LAI

680

720

550

700

Late Avar I!

LA II

720

760

700

850

Late Avar III

LA III

760

822

850

1000

Phase

Abbreviation

Early Avar I

..

--r
i
i

_,_I
lOOOAD

Sequenee- Sequeneedatesl
dates2

llOOAD

The archaeology of"ethnic groups" in the Avar qaganate

lee dates from

The correspondence analysis of burial assemblages with female skel

etons shown in Figure 8 resulted in two parabola-shaped distributions
joined in a single curve at the center of the graph. Since the chrono
logical sequence goes from the left to the right of Figure 8, the two
parabola-shaped distributions are to be dated to the Early Avar period.
A elose examination ofboth distributions indicated the upper parabola
consists of assemblages with artifacts viewed as "Germanic," while the
lower parabola indudes assemblages with "Byzantine-Avar" artifacts. In
both cases, the labels attached to such artifacts are based primarily on
the evaluation of analogies found for most of these artifacts in pre-Avar
assemblages in the Carpathian Basin or contemporary assemblages in
Central and Western Europe, in Italy or in the Balkans. Whatever their
narnes, the two distinct parabolas suggest that during the Early Avar
period, "Germanic" wornen were distinguished in dress from "Avar"
wornen wearing mostly dress accessories of Byzantine origin. By the
Middle Avar period, that distinction disappeared, as a consequence of a
drarnatic blending of traditions, and no such distinctions existed during
the Late Avar period. If there is any need oflabels for that period, then
the most recent assemblages on the right side of the graph could easily
pass for "Slavic" graves .
Besides chronology, chorology is of great importance for deciphering
and "reading" the material culture of the Avar age. With WinSerion,

age. The shift

rhich previous
u, while most
ufactured and
. of the person
,f the burial is
:quisition that
'cl. the shift to
~ undermines
le coexistence
le Carpathian
"'Avaria" must
'en though no
ginning of the
chronology of
seriation with
. new dates in
sAD 2:'

l,i

-~~,

60

PETER STADLER

functional and archaeological artifact categories were mapped sepa

rately. Over 7,000 maps were thus generated, onlya few of which will be
presented and discussed in the remaining part of this chapter. Given that
no archaeologist is capable of evaluating that many maps at the same
time, I developed and employed the "analysis of the N next neighbors"
precisely for facilitating the understanding of all map distribution con
sidered at any point in research. 1O The method allows, for example, the
concomitant evaluation of thousands of Avar-age ceramic pots depos
ited in Middle or Late Avar graves, all in a single map. The result of
that analysis delineate fourteen clusters, which may weIl be just as many
different settlement areas, within which trade seems to have been more
intense than with other areas. Such clusters could of course be checked
for other diagnostic artifacts, such as Late Avar casts. The spatial distri
bution ofthe fourteen clusters is shown in Figure 19.
What such maps can certainly show is not only how many different
settlement areas there were in the Avar qaganate, but also that that pol
ity was by no means homogeneous from a cultural point of view. In
other words, and pace Istvan Bana, Avars were most likely not the only
inhabitants of the Avar qaganate. Whether settlement areas identified by
me ans of the "analysis ofthe N next neighbors" could be further equated
with more or less known ethnic groups within the qaganate, is of course
a possible, albeit by no means unique, interpretation. Equally significant
is the mapping of functional types within one and the same cemetery in
order to identify spatial clusters possibly associated with the use of that
cemetery by different groups.
The tendency among archaeologists and historians is to treat the
culture of the Avar age as uniform, especially during the Middle and
Late Avar periods. However, a careful examination of the archaeological
re cord reveals many local and regional variants. Regional variants are
particularly difficult to interpret in historical terms. When taking into
consideration several other sets of data, from written sources to anthro
pological information and natural resources available in any given area,
it becomes clear that while it may be possible in certain cases to identify
10 This method is not to be confounded with the statistical method known by the
same name. My method is based on checking map distributions by means of a statisti
cal test, to see whether or not distributions are random. Non-random distributions are
then included in the matrix of assemblages, which is again evaluated by means of CA.
The resulting eigenvectors are then subjected to a mono-variate cluster analysis. The
obtained clusters are again mapped on a combined distribution map for all investigated
maps of individual characteristics.

"work:
excIud
bly lin
stern ,

theaI1

offem
cautio
in a D
tance
cal aJI

patter!

SUf\i1i

"miDi

andSllj

bef~

..:

AVAR CHRONOLOGY REVISITED

~d sepa
h will be
lven that
he same
19hbors"
Ion con
lple, the
s depos
result of
asmany
enmore
checked
al distri

different
that pol
view. In
the only
ltified by
equated
)f course
!9-lificant
letery in
;e ofthat

treat the
Idle and
x>logical
iants are
mg into
,anthroTen area,
identify
"''Il by the

'a statisti
utions are
IIlS ofCA.
Llvsis.lhe
\'"estigated

61

"workshops" in metalwork or the production of pottery, one should not

exclude the possibility of strong commitment to local traditions, possi
bly linked to groups of immigrants. Archaeologists have now received a
stern warning about the misuse or abuse of the ethnic interpretation of
the archaeological record, while sociologically-minded historians have
offered alternative directions of research. 1l However, irrespective of all
cautionary tales, the distribution and combination of artifact category
in a manner as precise as possible remains a task of outmost impor
tance for modern archaeology. The use of large databases and statisti
cal analysis allows now a much more refined understanding of cultural
patterning than previously possible. It i8 of course just as clear that the
surviving archaeological record contains only a small portion of the
"living" culture at any given moment in time. Language, songs, gestures,
and so many realia that did not survive in the archaeological record, will
be forever irretrievable by archaeological means.
Archaeologists can nevertheless recognize cultural patterns and dis
tinguish between groups on the basis of combinations of cultural ele
ments. Any discussion about how such patterns and groups should be
interpreted must start with functional types, namely with the distribu
tion of artifact categories for whieh distinct functions may be asserted.
Such categories are stored in the Image Database "Montelius" in the
field "TypOI:' Sometimes adjustments needed to be done "by hand" if
diagnostie characteristics were obscured by too large a elassification.
For example, "lance" proved to be too general; instead, more narrowly
defined types, such as "spear;' "leaf-shaped lance;' and "winged lance"
had to be taken consideration. It goes without saying that extra caution
is therefore needed in the interpretation, for occupational groups could
easily be mistaken for "ethnic groups:' A cluster of burial assemblages
with winged lances is not necessarily an indication of a group ofFranks,
but is certainly an indication of a group of specialized warriors.
The matrix showing the incidences of assemblages and functional
types was then subjected to a seriation by correspondence analysis. 12
Recurrent artifact categories with more than 500 occurrences were
eliminated for computational reasons. ather categories, such as iron
buckles, were regarded as without any diagnostic potential and were
likewise exeluded. The scattergram in Figure 11 shows the result of

" Brather 2004; Geary 2002. See also Curta 2007, with a critique ofBrather.
For the software, see WinSerion homepage at http://www.winserion.orgi.

12

:,~'

0\

4NRZ

AUCl''lLtJalt:>

A. R".i1erkarnpl

&.

Naa.1

..tt..HRBGreif

.~:tn(>;lVlfllPf

. . . Schtli'llle{,reil

MoIIUsk

.A. Vll yelnade4

Flng.1.7~!tt~nlPUnl
RZKelteozierPun;
Ohrrhk1AU

A lIIRB~ Anhanger
f
.Al'
HRIHlerkatTlp

A Ge!~Kerarnik

.At&.. Rieme,.".nhtmger

"Phdl~n'

Az..umLS

laumB~'Kid"9Ft'rau

iI.

Siirnbpschlag

...nt'fkl1(j~n _
,Di.mem
...
&~rnnqPerMl1h
AUalve1f

OhrrmgKt.lgi!k..hen

"0

i"he'

AGOrtel

.4 Schlo

~nh"r~rl*!4

..&.

.A,RaSs.e!

AAUf"'u

.a.stjrnbu!.Chhalter

(')hrrrngr.u-nuja_k:>t1~natt'

A ,,",H\Jf(~I$i'1n
.

..

.At..

tT1
!:>O

~g" "'''PI".I
ame

Schel

a-\, . .

AA

'"s;!

alv)/;:na er

t:1
AHRl

! __

ru:1k.r

..&.

lopfspange

Pir\lJ:rtte
S"iiAdegerat

t-<

"'JIlIU,iI.
Press.m.tHk'j..." Gewi(ht

YliJ~~<fi:h
"'lIt
A

~mmlR

~~k,n91'BTamcp W<l(lge
.T"~hlil9

S<hoaUela
A Anh'ng<eSpi

AHefttfl1

lt,'RBMill. . . RZfamga

~ild

tT1
!:>O

. . .T~ BesrnldgZa
AAttach'"

Spatha
RZFe

J/iont.iftlitatton

..

Kd;erbe.chl.~g

~ Boget'l~Khla9

Figure 11. Zoomed detail of the correspondence analysis scattergram of functional types of artifacts from
Avar-age burial assemblages.

',"&4~",~,,-,l

S1:&2

:;:~ ~ ~

::l.&

;s..PlO:n>

~ llf~hJ'!-.l~_

i..Q

&..i RJi lieS l,l.h..R M"

,; 'I&A.--o

.',2.

pA . . . . . . . . . . .

AVAR CHRONOLOGY REVISITED

eo

..t::

'"
u

t;

o;j

'0
<J)

Q.I

E
--;
s::

"B

t::

..El
......

e .

~ ~

e.o~

v.D

:: e
<tf

<l)

:Q

o;j

~ ~

",-

r~
o;j ...

;.E

<l)

<l)

...

<tf

"

o;j

OJ)

~
0..

'"

t
o

.s......o

63

the correspondence analysis of burial assemblages with both male and

female skeletons. One of the most evident conclusions is the separation
of functional types into two gender-specific sets. In the space for factors
2 and 3, the correspondence analysis shows a clustering of functional
types, which can only be interpreted in terms of gender. The upper right
corner ofthe scattergram is occupied by artifacts most commonly found
with female burials, while the lower left corner is reserved for artifacts
usually found with male burials. Between these areas of the scattergram
are those functional types which are not gender-specific. Functional
types are displayed in such a way, that symbols shown next to each other
(sometimes even overlapping) correspond to artifact categories that fre
quently appear together in the burial assemblages. The more frequently
functional types appear together, the larger the symbols used on the
scattergram .
The following six clusters can be identified for assemblages with male
skeletons:

Cluster 1: quiver mounts, bone reinforcement plates for the reflex

bow, plait clasps, (earrings in male graves) .
Cluster 2: bag fasteners, bone mouthpiece for drinking horn, T-shaped
mounts, bone or antler instrument for untying knots, saber, armor
plates, single- and double-edged swords.
Cluster 3: spatha,13 sax,14 tweezers, shield bosses, helmets, belt fasten
ers for the buckle.

Cluster 4: plowshares, scythes, chisels, sickles, horseshoes.

Cluster 5: bridle hole guards, bridle strap pendants, bridle forehead

mane lock holders, phalerae, bridle forehead mounts, (stirrup, snaffle) .
Cluster 6: blacksmith tools, anvil, rasp .

'c;j

<l)

"
"

Five more clusters have been identified for assemblages with female
skeletons:

N
,...;
~

;:l
OJ)

i.i:

Ll The weapon known to archaeologists of the early Middle Ages as spatha goes back
to the first century or perhaps to similar weapons of the Latene tradition of the last cen
turies B.C. The typical weapon of the Roman legionnaire, the spatha was a straight, 0.75
to 1.0 m long, doubIe-edged sword with a Iong tip. As such, the spatha is much broader
than either single- or double-edged Avar-age swords found in horseman burial assem
blages. The latter later developed into the Middle and Late Avar sabers.
14 A sax (also known as scramasax) i5 a single-edged, long knlfe.

64

PETER STADLER

Cluster 7: "Slavic" bow fibulae, spiral pendants, weights, scales, keys,

chains, and combs;
Cluster 8: choppers, "Germanic" bow fibulae, garter sets for leg bind
ings, crosses, casket mounts, belt buckles with dentil ornamentation
(Zahnschnitt), pendants.
Cluster9: strapends withdentilornamentation (Zahnschnitt), T-shaped
mounts, belt mounts and bracelets, "Merovingian" pendant set.
Cluster 10: hair ornaments, diadems.
Cluster 11: earrings with spiraled pendants.
Table 2: Classification of clusters by "ethnic groups"
Group

Male

Female

Horses

'~var"

ClusterOl,
Ciuster02

Cluster I 0

Cluster05

"Germanie"

Cluster03,
Cluster06?

Cluster08,
Cluster09

"Slavic"
"Byzantine"

Clusterll
Cluster06?

Cluster07

Cluster 4 indudes mostly agricultural tools and implements which

appear more often in hoards than in burial assemblages. 15 Similarly,
Cluster 6 indudes mostly blacksmithing tools and could thus be attrib
uted to another "occupational group;' namely that of craftsman buri
als. 16 That Cluster 6 is dose to both the "Germanic" Cluster 3 and the
"Byzantine" Cluster 7 may indicate that no ethnically specific attributes
were linked to the social status associated with craftsmen in Avar society.
Much ink has been so far spilled on the presence of artifacts of Byzan
tine origin in Avar-age burial assemblages. Dezs Csallany was among
the first to call attention upon the so-called "Byzantine" belt buckles, a
line of research now continued by Ursula Ibler and Vladimir Varsik. 17 In
arecent monograph, Eva Garam has gathered in its entirety all artifacts
found in Avar-age burials, which have been regarded as of Byzantine
origin. 18 While her work deals primarily with the Early and Middle Avar

Curta 1998-1999.

'6 See Orsolya Heinrich-Tamaska's contribution to this volume.

17 Csallany 1954; Ibler 1992; Varsik 1992.

'8 Garam 2001.

l5

I
i
I
~.

of

.,
:lJ

:ales, keys,

r leg bind
mentation

"T-shaped
et.

,rses

iterOS

~nts

which
Similarly,
s be attrib
iman buri
. 3 and the
: attributes
rar society.
ofByzan
ras among
buckles, a
larsikY In
II artifacts
Byzantine
tddle Avar

AVAR CHRONOLOGY REVISITED

65

periods, Falko Daim has recently analyzed a group of Late Avar belt
buckles and mounts to which he attributed a Byzantine origin. 19 What
all those studies have shown is that most artifacts regarded as Byzantine
were most likely imports and are therefore not necessarily an indication
of the presence within '~varia" of a Byzantine population. 20
The fact that bow fibulae which Joachim Werner first called "Slavic"
appear in Cluster 7 together with other "Byzantine" artifact categories
seems to confirm the conclusions of Florin Curta's studies, whieh have
meanwhile raised serious doubts about regarding such fibulae as badges
of Slavic ethnie identity.21 On the other hand, there can be no doubt
about the presence of the Slavs inside the qaganate, whieh is weH docu
mented in written sources. But there are apparently no "Slavic" artifact
categories, an indieation of the low resolution at whieh labels of "ethnie
groups" have so far been used in Avar archaeology. The famous lock or
ear-rings with S-shaped twisted end may well be a chronologically spe
cific artifact category, given that such rings appear at the end of the Avar
chronology, but continued to occur in post-Avar assemblages dated to
the ninth century long viewed as "Slavic:'22 I shaH return shortly to the
problem of the Avar-age Slavs. Meanwhile, a number of burial aspects,
such as inhumations with tunnel-shaped shafts, have been cited for eth
nie attribution, but work on this part of the database is still in progress. 23
Until then, the attribution of such graves to groups of nomads from the
steppes north of the Black Sea (Bulgars or Cutrigurs) may be treated
with caution. Cluster 3, 8 and 9, which can be assigned to "Germanic
tribes': lead over to the following section.

7he "Germanie" population of the Avar qaganate

Clusters 3, 8, and 9 have been tentatively labeled "Germanic" because
of the artifact categories used for their definition. For a long time, most

Daim 2000.
BaHnt 1983 advanced thc idea that the cluster of"Byzantine" artifacts in southwest
crn Hungary, in thc region of the Balaton Lake and around Pecs, may signal the presence
of the Sermesianoi mentioned in the Miracles ofSt. Demetrius.
21 Werner 1950 and 1960; Curta 1994, 2004, 2005, and 2006; Curta and Dupoi 1994
1995.
22 Tbe idca that thc lock ring with S-shaped end is "Slavic" gocs back to Lubor Nie
derle and is weil entrenched in the archaeology of the early medieval Central Europe
since Eisner 1933 and Korosec 1951.
23 For inhumations with tunnel-shaped shafts, see Lrinczy 1994 and 1995.
19

20

66

PETER STADLER

Hungarian archaeologists rejected the idea that any Germanic groups

may have existed within the Avar qaganate. They argued instead that
in 568, with the departure of the Lombards to Italy, all Germanic ele
ments had moved away leaving Pannonia completely deserted. The most
articulate advocate of such a theory was Istvan B6na, whose ideas must
be viewed as areaction to the ethno-chronological interpretations of
Joachim Werner. 24 On the basis of a cavalier treatment ofthe Varpalota
cemetery, Werner believed that since "Lombard" and "Avar" graves in
that cemetery were found side by side, not all Lombards had taken off to
Italy in 568. B6na rightly retorted that the "Lombard" and "Avar" burials
in Varpalota were not coeval and that a relatively long period of time
separated the ones from the others. However, with his reaction B6na
threw the baby out together with the bathwater. He began rejecting any
arguments, valid or not, pertaining to Germanic cultural elements of
the Avar age. A widely recognized authority on the archaeology of the
early Middle Ages, both in his country and abroad, B6na silenced any
opinions that contradicted his theory. This may explain why his former
student Gabor Kiss was able to write an excellent study of the earrings
with mounted bead in Pannonia, without any reference whatsoever to
their ethnic attribution. 25
Archaeological excavations in the late 1960s and 1970s produced
even more evidence of "Germanic" cultural elements in the Transda
nubian region of Hungary. As a consequence, Attila Kiss proposed that
after their defeat by the Avars, large groups of Gepids were forcefully
moved to Pannonia. 26 Nevertheless, the problem may now be revisited
in the light ofan ever increasing number of finds. The excavation oflarge
cemeteries such as Krnye, Klked Feketekapu A and B, Zamardi, and
Budapest-Budakalasz has produced sufficient evidence to demonstrate
that after the Avar conquest of 568, "Germanic" cultural elements not
only survived but also developed in direct contact with the Merovin
gian world. This points to a certain prosperity during the Avar age of
a relatively large population, which the Avars had found in Pannonia.
Cemetery A in Klked Feketekapu began most likely in the aftermath
24 B6na 1971 and 2000. B6na ignored Werner's studies published after his book on
Lombards in Pannonia (Werner 1962). His only direct comments on Werner's ideas
about the Varpalota cemetery may be found in B6na 1971, 301, but in reference to Dezs
Simonyi.
Kiss 1983. Earrings with mounted beads are now seen as "Germanie:'
26 Attila Kiss first presented his ideas in 1979 (Kiss 1979). See also Kiss 1984,1987,
and 1996. For cemetery B, including an aristocratic female burial attributed to a Gepid
lady, see Kiss 200 L

AVAR CHRONOLOGY REVISITED

groups
!itead that
l1anic ele
Themost
:ieas must
tations of
Varpalota
graves in
tken offto
lr" burials
Id of time
tion B6na
~cting any
ements of
)gy of the
enced any
lis former
e earrings
tsoever to
je

produced
. Transda
>Osed that
forcefully
e revisited
)noflarge
lardi, and
monstrate
nents not
Merovin
~'ar age of
Pannonia.
aftermath
I1is book on
rner's ideas
Iceto Dezs

67

of the Avar conquest. Some time after the local community began bury
ing its dead, an ''Avar governor" was also buried on the outskirts of the
graveyard, together with his wife and child. During the first occupa
tion phase, until about 580 or 590, the burial of the "Avar governor"
was the only connection to ''Avaria'' of the "Germanic" community in
Klked Feketekapu. A population of different origin and conspicuous
Avar culture began settling among the natives only after that. The new
comers opened ground for a different cemetery (cemetery B), in which
there is clear evidence of a blending of cultural traditions. Cemetery
B ends at some point in the 600s, after which occupation ceased com
pletely. A new occupation occurred only in the 700s, when members of
yet another group settled in Klked Feketekapu. By that time, aH "Ger
manie" cultural elements had disappeared without any trace. The third
occupation phase in Klked Feketekapu is therefore characterized by
the "standardized" culture of the Late Avar period.
But what were the cultural differences between "Germanic" and ''Avar''
burials? As mentioned before, the main distinctions are to be drawn in
clothing and weapons. Combs, belt sets ornamented with dentil orna
mentation (Zahnschnitt), spathae or short dagger-like swords known
as sax appear only in "Germanic" burials. By contrast, gold earrings
(which appear in burials ofboth males and females), plait clasps, quivers
and bow bone reinforcement plates, single-e and double-edged swords
with P-shaped attachments are all typical for "Avar" burials. Whatever
the ethnic identity of those burying their dead in "Germanic" graves,
the evidence from the two cemeteries excavated in Klked Feketekapu
clearly points to sharp distinctions in material culture, which may have
weH marked ethnic boundaries. There are several ways in which this sit
uation may be explained historically. The "Germanic" cultural elements
may indieated the presence of a Lombard group that did not migrate
to Italy; of a Gepid group forcefully resettled from the eastern regions
of the Carpathian Basin; of a group of Sueves who had survived under
Lombard and now under Avar rule; a mixture of all these groups, as weH
as others not mentioned in the written sources.
During the last few years, Hungarian archaeologists excavated the
until now largest Avar-age cemetery in Zamardi, on the shore of Lake
Balaton. 27 Zamardi stands out among all other contemporary cemeter
ies by means of the large number of graves so far revealed (about 6,000)
and the conspicuous prosperity of the Avar-age community burying its

1984, 1987,

I to a Gepid
For a preliminary report, see Bardos 2000.

68

PETER STADLER

dead in that cemetery, which is evident in the quantity of gold and sil
ver belt sets recuperated from otherwise extensively robbed burials. On
the basis of both the size and the wealth of the cemetery, Istvan B6na
even suggested that Zamardi must have been a center of Avar power, an
ordu. 28 He saw no contradiction between such an idea and the fact that
most belt sets found in Zamardi have a dentil ornamentation (Zahn
schnitt) most typical for "Germanic" assemblages and evidently inspired
by the tradition of the Animal Style I. Equally interesting are the good
analogies in the western and southern Merovingian regions that can be
:= ]
established for belt buckles and mounts used to decorated shoe laces
or for belt-shaped pendants found in female burials. The evidence in
any case bespeaks the considerable wealth of a group, possibIy of Ger
manie origin, wh ich throughout the Early Avar period maintained dose
relations with distant communities in southern Germany and France.
The wealth of the Zamardi community may perhaps be attributed to
the participation of its members in the Avar campaigns against the early
Byzantine Empire.
"Germanic" traits have a peculiar geographie distribution. Figures 12
and 13 show the cluster of belt sets with dentil ornamentation (Zahn
schnitt) in Transdanubia. 29 The dentil ornament is currently regarded as
a Iocal development of the Animal Style II post-dating the conquest of
Pannonia by the Avars. The duster of finds in Transdanubia may indi
cate that this style of decoration originated from the lands on the shores
of Lake Balaton, which had been under Lombard control before 568,
even though artifacts with dentil ornamentation have also been found
along the Tisza River in formerly Gepid territory.
Two other maps (Figs. 14 and 15) showthe distribution ofthe archae
ologically attested custom of the comb deposition in graves. Attila Kiss's
excavations in cemetery A at Klked Feketekapu revealed that in both
~=
male and female burials combs often appear either on the Ieft or the right
~ ~,
side of the skulI, which suggests that they were perhaps meant to look
j'" ,
as if worn in lifetime. The distribution of graves with combs overlaps
~~
that of dress accessories with dentil ornamentation, even if, because of
i~
the specific state of research, the comb finds from cemetery A in Klked
Feketekapu seem to dominate the picture. Combs and dress accessories
;..~
with dentil ornament appear especially in those areas, wh ich before 568
~-

.c J
:1
:c

~i

28
29

Istvan B6na, personal communication, 1990.

Heinrich-Tamaska 2007.

<;.;I

Vi

Q)

<

rt>
i 1 i:5 0: '"J:rt> ......
,.., O'::!.
~ ::1. ~ o. o OQ
......,,,,
~ g'
oo~~

Ei

Q)

0" ~ ::J"'
0 '" P'
~" rt>I

s-

':i

;;,

,~

I:

a"Nf:;

::J"' ..... rt> rt> s:l ~

:::
J: V1 o
p... ~ p...~ ......
::: 0\ ,..,
p...S-O rt> ...., 0
......, 'P'" ::!I N

'"

iil
rt>

~J

~
p... ~

(l..

n..
0

rt>
~ Ort>'"
....

(I

;:;..,
Cl

::;

(l>

(I

......

,..,rt>, ~. ~
::: '"

(')
f"11
rn ~. (I ~
.... rt>
Q) OQ
"
:::
0 ' t:;.
~
0" 0 rt> ::!
rt>p...p... , ~ :::

I=-:::J
t;:j:np...
[!l.

0
S-~ ,
Pol :::

.,..

Orsolya Tama,ka, Peter Stadler 2006

Zahn".;;chnitt im Karpatenbecken

~
;;
:xl
()

::r:

z
ot-'
oQ

g;
:5Cfl

::i

tT:I

ZahnschnittA Tierornamentik Gruppe A

.ZahnschnitiH Tierornalnentik Gruppe B
", ZahnschniHC klassisches Flechtband

.. ZahnschnittD awarisches Flechthand

I ZahnschnittE Tierdetails
'ZahnschnittF figural

Figure 12. Distribution map of dress accessories with dentil decoration

Data after Heinrich-Tamaska 2007.

0\
'-0

ornamental motifs.

Tama~ka,

Peter Stadler 2006

'-J

"0

t'J:j
...,
t'J:j

:;.;;

'"~

ot"'
t"rj

:;.;;

...
,..
....
I

MaterHerstAa Gold gegossen und getrieben

MaterHerstAb Gold gepresst
MaterHerstHa Silher gegossen
MaterHerstb Silber gegossen vergoldet
MaterHerstBc Silber gepresst

MaterHerslCa Brot1u: gegossen

MaterHerstCb Bronze gegossen vergoldet
.. MalcrHcrslCc Bronze gepresst
~

Figure 13. Distribution map of dress accessories with dentil decoration

Data after Heinrich-Tamaska 2007.

"kIlU'lk

1'11I11I1I'... r:N,xINL'iN

techniques.

'111 l;otllNivN :; NllI'lllknor- 0 1:r<"luOn

1uIl......~ _ Ir'

 Ml1klllrrlltAh (;01<1 Mrpn'!tfld

,. M,llt"rI kn,tHlJ Sdbt'r ~(.'go!\scn
... MaterHerstBb Silber gegossen vergoldet
MaterHerstBc Silber gepresst

.. Mnt('tf Irr;'!t(:h nruH7r l,(t.'l,(Wlllt'1I v~l'f.t0hlt't

- MJ.tert !erst(:l Bronze g-l.'pn'sht

MaterHerstCd Bronze Ileoresst

Figure l3. Distribution map of dress accessories with dentil decoration by production techniques.

Data after Heinrich-Tamaska 2007.

aktuelle Parameter:NextNeiN=90 KonfNivN=5 Normkoor=O Frequ=On

~
>
~
C"l

::x::

~
z
ot""'
o
Cl

><

.....
(/J
.....
'"'I
tr.I

t:I

Kamm 1 - reihig
N= 25 [43], wirklicher M.w.= 1.83, erwarteter M.w.= 0.80, Differenz= 1.04, Konfidenzniveau= 100.0%

= Ix,

=2x,

=3x,

=4x

Figure 14. Distribution map of combs with teeth in a single-row deposited in graves.

'-l

.......

aktuelle

Parameter:NextNeiN~90

KonfNivN-5

N()nnko()r~O

I'requ=On

'-l
N

"C

tT1

'"l

tT1

~
[J)

I:)

t-<
tl:I
~

Kamm 2 - reihlg
53 (211], wirklicher M.w.~ 3.07, erwarteter M.w.; 1.69, Differenz; 1.38, KonfiJenzniveau; 100.0%
.=<15x,
~<30x,
=<45x,
=<60x,
=<75x,
~<85x

N~

Figure 15. Distribution map of combs with teeth in a double-row deposited in graves .

D"t

tI'

~i~I

AVAR CHRONOLOGY REVISITED

<J)

73

were inhabited by Lombards and Gepids, respectively. This is of course

not to say that responsible for the phenomenon must only be Lombards
and Gepids surviving under Avar rule. It may well be that other groups
within the qaganate adopted those cultural traits. But their distribu
tion is quite distinct from other cultural traits which have been labeled
"Slavic" (in the northwestern region of the qaganate), "Romance" (at the
southwestern tip of Lake Balaton, the so-called "Keszthely culture:' or
"Byzantine:'30

bh
.S
"Cl

.~
o
<J)

Ihe interpretation of the "Slavic" assemblages in the

northwestern region ofAvaria

0...

~
~

~
'll

'""o"

"Cl

'"

"'""
.~ ~
E~
11

.S'"

-=
CJ

~. -B
~
00
~
11

2l

'"

.~

<J)

><
~

~e

es

..0

S
0

0...

d:
>D

. ><

Ei'"

'"
~ t i:1
.....;:l
;;;:. .S
~

~ ~
"

o.r;
><

'" ..,.v

"~."
:<"

;;;:
~ ><
.c",
'" '"
~ t

:.0'=

..0

'C
.....
<J)

i5
lf)

,....,

5b

Figure 16 shows the distribution of ceramic pots found in Middle and

Late Avar assemblages. There are of course a few clusters, but all in all
the deposition of ceramic pots in graves was a wide-spread phenom
enon. By contrast, Figure 17 shows the distribution of ceramic wares
with prick-like comb punch decoration (Kammstich). The distribution
is remarkably similar to that of wares with potter's marks on the bot
tom of the pot. 31 A combination of all traits pertaining to ceramie wares
by means of the analysis of N next neighbors pro duces the distribution
map shown in Figure 19, on which wares with prick-like combed punch
decoration and potter's marks appear as clearly distinct clusters in the
northwestern area of the Carpathian Basin (groups 9-12).
At a close examination of the history of settlement in the north
western region of Avaria, it appears that a substantial occupation of
the region only began in the early seventh century, ca. 630. Both seria
tion and radiocarbon dating confirm that the northwestern region
was settled at about the same time as the northeastern region on the
Upper Tisza. In the northwest, burial assemblages with wares deco

rated with prick-like comb punches and potter's marks are attested

throughout the Middle and Late Avar period, from ca. 630 to ca. 800. In

other words, throughout much of the Avar age, such traits as prick-like
comb punch es and potter's marks seem to have typical primarily for the

-=
"
"E ::5
!",l~><

"'::C:
o.r; V
:-:
:.L::

11

11

Z ..

30 For the Keszthely culture, see Kovrig 1958, Kiss 1967, Mller 1996a and 1996b,
Bierbrauer 2004.
31 As this is a much debated topic in the archaeology of medieval Eastern Europe, an
abundant literature exists on the topic. Only a few, most important titles may be cited
here: Com~a 1961 and 1973; Diaconu 1986; Kolos-Szafranska 1953; and ToCik 1962.

--

.'..

~.

_-",:c> __

j,'

;~ " ;~~;;i'r"'\
,'6c.,
'i';,
.. . ' II/f"
","

/"

;;;. ,i"

'

/;

,.

;;t;;
"",

~'O" .~,o,'
' -'''"'''''''''.
~;?r e. '.',.
"""""":~",.,,-.f" ,';"

.i,.J

~./
0
,
,

... ...."

;\~;Jh::;.-,;,\.,\ \

,.

.;.

."

~
...,

,,~

",--I'!i
"', ..
'c:......
fII
f!',
~31
., '
',#.''--L..,
" .,
.',

..

'

;:0
Vl

;;

""'i

..~_
. ' ~l!{
'/.'.!i;.,'
;;
,"
i

tT1

,....

"

".;"~

'

ot'"'
tTl
;:0

.'

Keramic Topf Funktioneller Typ

N = 562 112501]

=<237x,

=<474x,

=<711x,

=<948x,

=<1l85x,

=<1422x

Figure 16. Distribution map of ceramic wares in the Carpathian Basin.

 =<237x,

=<474x,

~<711x,

,,-<948x,

=<1185x,

=<1 422x

16. Distribution map of ceramic wares in the Carpathian Basin.

:>

::t
~

o
Z
o
r
o

Cl

.-<:

g;

<':

l:i>

::J
!:Tl
o

...
.,
...
11

Topf00240
Topf00250
TopfO0270
TopfO0280
TopfO0290

Kerall1iktl('pfiverziertiKammstempel
Kerall1iktlbpllvcrzicrtlKamll1stempel/Strichreihe
Kerall1iktlbpllvcrzicrt/Kammstempel/Fingernagel
Keramik/'lbpf/verzicrt/Kammstempel/Fischgrt
KeramikiTopf/verzicrt/Kammstempel/Kall1mstempel

.. '1()pfO03 JO
... Topf00320
'" Topf00330
"" TopfO0310

Keramik/Topflverzicrt/Kammstempel/Punktreihe
Keramik/Toptlverziert/Kammstempel/S
Keramik/Toptiverziert/Kammstempel/Stempel
Keramik/Topf/verziert/Kammstempel/Strichimitation

Figurc 17. Distribution map of ceramic wares with prick-like comb punch ornament.

'-.J
V1

'-l
0'\

'r;I

..,
tr1

tr1
:;.:I
(.I')

I:j

t""
tr1
:;.:I

....

'V

KcramikITopfiverziertiTpfermarke/Kreis/klein
Keramikl'l'opfiverziertrnipfermarkeIKrcuz/

Figure 18. Distribution map of ceramic wares with potter's marks.

 '1"l'm()~70 :
'l 'lop(00580 : KeramikrJbpflverzkrtlTpfermarkelKreuzl

.. '1(11'11)(1050 : K,'n\lllikl'ltll'flverzk'rll'mplrrmarkdlhllll'l1

Figure 18, Distribution map of ceramic wares with pottds marks.

>'

:0
(')

:J:

:0

o
Z

ot-<
o

Cl

g;

:s.....

(J)

>-l
tTl

tl

..

T
....

CluOl
Clu02
Clu03
Clu04
Clu05

..

T
....

Clu06
Clu07
CltlDS
Clu09
ClulO

..

T
...

Clull
Clu12
CluLl
Clu14
I ClulS

..

T
...

Clu16
Clu17
Cl"IB
Clu19
Clu20

(;) Clu21
A Clu22
VI CluB
-<l Clu24
m Clu25

o Clu26
A Clu27
'V Clu2S
-<l Clu29
Clu3U

..
A
T
...

Clu31
Clu32
Clu33
Clu34
I CI,,35

GI

T
...

Clu36
Clu37
Clu38
Clu39
Clu40

'" Clu41
A Clu42
VI Clu43
<l Clu44
B Clu45

(l) Clu46
A Clu47
VI Clu48
<I Clu49
9 Clu50

Figure 19. Plotting of the analysis ofN next neighbors for all pottery features associated with Middle and
Late Avar burial assemblages,

"'-J
"'-J

78

PETER STADLER

northwestern region of the Avar qaganate. Outside the qaganate, such

traits appear only in the neighboring regions-the western and north
western parts of Lower Austria and Moravia-in which a massive pres
ence of the Slavs is often assumed for the century following the collapse
of the Avar qaganate. Within the qaganate, the only other, but much
smaller cluster of burial assemblages that produced wares decorated
with prick-like comb punches is in the environs of pecs. It becomes
therefore apparent that beginning with Middle Avar I a regional identity
may have formed in the northwestern lands of the qaganate, which was
marked in funerary contexts by means of both ritual and the deposi
tion in graves of ceramic wares with specific ornaments. It is quite pos
sible that the northwestern lands had been under Avar control since the
beginning, but no signs exist of a serious settlement before ca. 630. That
date remarkably coincides with the rise ofSamo's polity known from the
Chronicle ofFredegar. 32
A further indication of the special nature of the northwestern lands
of the Avar qaganate is the cluster in that region of the largest number
of warrior graves. This suggests a sudden military presence of the Avars
in the area, perhaps in the aftermath of Samo's rebellion. If the region
was part of Samos polity, it must have returned relatively quickly to Avar
rule, this time reinforced by the military posturing of the population
settled in the region. Indeed, the only area within the qaganate where
such a deliberate policy of settlement is so evident in the archaeological
record is the northwest. Avar-age burials, particularly horseman burials,
in the northwest seem to have been systematically robbed after ca. 800.
Whether or not this phenomenon may be attributed to the revolt of the
former Avar subjects, groups 9-12 in Figure 19 must be seen as a reac
tion to the particular political and military circumstances of the early
seventh century. That some ofthe cultural traits in those groups outlived
the Avar qaganate further suggests that that reaction resulted in invent
ing cultural traditions oflong-term political consequences.
The interpretation advanced in this chapter is based on a much
improved chronology, itself the result of refined methods combining
traditional seriation with radiocarbon dating. My only hope is that an
improved chronology may contribute to a new evaluation of the prob-

Fredegar 4.48, in Wallace-Hadrill1960, p. 40. For the chronicIe, see Goffart 1963,
Kusternig 1982, and Wood 1994. For Slavs in Fredegar, see Curta 1997. For a survey of
the abundant literat ure on Samo, see Eggers 2001.

Bilint..
tsa

AVAR CHRONOLOGY REVISITED

79

il1ate, such
md north
ssive pres
le collapse
but much
decorated
t becomes
Lai identity
which was
ne deposi
quite pos
1since the
.630. That
n from the

lern of ethnicity in the archaeology of the early Middle Ages. Sebastian

Brather's critique of traditional approaches has done much to advance
our awareness of the pitfalls of an archaeology of ethnicity. However,
he did not propose anything to replace the supposedly outdated mod
els. The very absence of any alternative is an indication that for Brather
ethnicity should be banned from the archaeological vocabulary. My
own understanding of the archaeological record avoids the pitfalls of
Brather's agnosticism and advocates instead for the use of refined meth
ods of establishing relative and absolute chronologies, as a preliminary,
but necessary phase in the study of cultural patterns that might, under
certain circumstances, mark ethnic boundaries .

tern lands
st number
fthe Avars
the region
rlyto Avar
x>pulation
late where
i3eological
an burials,
er ca. 800.
volt ofthe
las a reac
f the early
:tSoutlived
in invent

Balint, Cs. 1983, "Az avarkor es honfoglalaskor bizanci vonatkozasainak regeszeti kuta
tasa Magyarorszagon 1970-1980 kztt': Antik Tanulmanyok 30, 116-25.
Brdos, E. 2000, "La necropoli avara di Zamardi'~ In wm degli Avari, ed E. Arslan and
M. Buora, Milan, 76-143.
Bayes, T. R. 1763, '~.c\n Essay Towards Solving A Problem In The Doctrine Of Chances':
Philosophical Transactions of the Royal Society, 53, 370-418.
Benzecri, J. P. 1973, Ilmalyse des donnees. II: D:malyse des correspondances. Paris.
Bierbrauer, V. 2004, "Die Keszthelv-Kultur und die romanische Kontinuitt in Westun
garn (5.-8. Jh.). Neue berleg~ngen zu einem alten Problem': In Von Sachsen bis

[} a much
;ombining
is that an
'the prob

ioffart 1963,
ra survey of

References

Jerusalem. Menschen und Institutionen im Wandel der Zeit. Festschrift fr Wolfgang

Giese zum 65. Geburtstag, ed H. Seibert and G. Thoma, Munich, 51-72.
B0lviken, E. et al. 1982, "Correspondence analysis: An alternative to principal compo
nents': World Archaeology 14, 41-60.
B6na,1. 1971, "Ein Vierteljahrhundert der Vlkerwanderungszeitforschung in Ungarn
(1945-1969)': Acta Archaeologica Academiae Scientiarum Hungaricae 23,265-336.
~. 2000, "Ein frha""'arisches Grberfeld in der Unio-Sandgrube von Varpalota",
Communicationes Archaeologlcae Hungariae 2000, 123-60.
Brather, S. 2004, Ethnische Interpretationen in der frhgeschichtlichen Archologie
Geschichte. Grundlagen und Alternativen. Berlin/New York.
Bronk Ramsey, C. 2001, "Development of the radiocarbon program OxCal': Radiocar
bon 43 (2A) 355-63.
Com~a, M. 1961, "Cu privire la semnificatia mrcilor de olar din epoca feudal timpu
rie", Studii ~i cercetari de istorie veche 12, 291-305.
-.1973, ''L'origine des marques de potier sur la ceramique slave': In Berichte ber den
11. internationalen Kongre fr slawische Archologie. BerUn, 24.-28. August 1970, ed
J. Herrmann and K.-R Otto, III, Berlin, 167-73.
Csallany, D. 1954, "Pamiatniki vizantiiskogo metallobrabatyvaiushchego iskusstva r:
Acta Antiqua Academiae Scientiarum Hungaricae 2, 311-348.
Curta, F. 1994, "On the dating of the 'Vetel-Co~oveni' group of curved fibulae': Ephem
eris Napocensis 4, 233-65.
-.1997, "Slavs in Fredegar and Paul the Deacon: medievalgens or 'scourge ofGod'?':
Early Medieval Europe 6, no. 2,141-67.
- . 1998-1999, "Iron and potlatch: early medieval hoards of implements and weapons
in Eastern Europe;' Archivum Eurasiae Medii Aevi 10, 15-62.

80

PETER STADLER

- . 2004, "Werner's dass I H of 'Slavic' bow fibulae revisited'; Archaeologia Bulgarica

8, no. 1, 59-78.
-.2005, "Female dress and 'Slavic' bow fibulae in Greece'; Hesperia 74,101-46.

2006, "Slavic bow fibulae? Werner's dass I D revisited'; Acta Archaeologica Aca
demiae Scientiarum Hungaricae 57,486-535.
-.2007, "Some remarks on ethnicity in medieval archaeology'; Early Medieval Europe
15, 159-85.

Curta, F., and Dupoi, V. 1994-1995, "ber die Bgelfibel aus Pietroasele und ihre Ver
wandten", Dada 38-39, 217-38.
Daim, F. 2000, "'Byzantinische' Grtelgarnituren des 8. Jahrhunderts': In Die Awaren am

Rand der byzantinischen Welt. Studien zu Diplomatie, Handel und Technologietransfer

im Frhmittelalter, ed F. Daim, Vienna, 77-204.
Diaconu, P. 1986, "Wieder ber den Ursprung der Bodenzeichen auf Tongefassen': In
Zbornik posveten na Bosko BabiC. Melange Bosko Babic 1924-1984, ed M. Apostolski,
Prilep, 79-80.
Eggers, M. 2001, "Samo, 'der erste Knig der Slawen'. Ein kritischer Forschungsbericht",

Bohemia 42, 62-83.

Eisner, J. 1933. Slovensko v praveku. Prague.
Garam, E. 2001, Funde byzantinischer Herkunft in der Awarenzeit vom Ende des 6. bis
zum Ende des 7. Jahrhunderts. Budapest.
Geary, P. 2002, 1he Myth ofNations. 1he Medieval Origins ofEurope. Princeton.
Goffart, W. 1963, "The Fredegar problem reconsidered'; Speculum 18,206-41.
Greenacre, M. 1984, 1heory and Applications of Correspondence Analysis. London/
Orlando.
Heinrich-Tamaska, O. 2007, "Tier- und Zahnschnittornamentik im awarenzeitlichen
Karpatenbecken", Bericht der Rmisch-Germanischen Kommission 88, forthcoming.
Hines, J., Nielsen K, and Siegmund, F. (eds) 1999, 1he Pace ofChange: Studies in Early

Medieval Chronology, Oxford.

Ibler, U. 1992, "Pannonische Grtelschnallen des spten 6. und 7. Jahrhunderts';

Arheoloski Vestnik 43, 135-48.

Kiss, A. 1968, '1\ Keszthely-kultura helye a pannoniai r6mai kontinuitas kerdesben,"

Archaeologiai Erteslto 95, 93-101.

- - . 1979, "Das Grberfeld und die Siedlung der awarenzeitlichen germanischen Bev
lkerung von Klked;' Folia archaeologica 30, 185-92.
- . 1984, "Die Keramik aus dem awarenzeitlichen Grberfeld A von Klked-Feket
akapu, Ungarn", Archaeologia Austriaca 68, 335-38.
1987, "Das Weiterleben der Gepiden in der Awarenzeit': In Die Vlker Sdosteuro
pas im 6. bis 8. Jahrhundert, ed B. Hnsel, Berlin, 203-18.
1992, "Germanen im awarenzeitlichen Karpatenbecken". In Awarenforschungen, ed
F. Daim, I, Vienna, 35-134.
- . 1996, Das awarenzeitliche gepidische Grberfeld von Klked Feketekapu A. Inns
bruck.
-.2001, Das awarenzeitliche Grberfeld in Klked-Feketekapu B. Budapest.
Kiss, G. 1983, "Egy avarkori flbevaI6tlpusr61", Archaeologiai Ertesit6 110, 100-11.
Kolos-Szafranska, Z. 1953, "z badan nad znakami gamcarskimi z terenu ziem polskich",
Wiadomosd Archeologiczrle 19, 180-97.
Korosec, J. 1951, "Delitev Siovanskih kultur zgodnjega srednjega veka v Jugoslaviji:'
Arheoloski vestnik 2, 134-55.
Kovrig, L 1958, "Megjegyzesek a keszthely-kultura kerdesehez'; Archaeologiai Ertesit6
85,66-74.
- . 1963, Das awarenzeitliche Grberfeld von Alattyan. Budapest.
Kusternig, A. 1982, "Einleitung". In Quellen zur Geschichte des 7. und 8. Jahrhunderts, ed
H. Wolfram, Darmstadt, 3-43.

Figures

L I
Cl

AVAR CHRONOLOGY REVISITED

81

Lrinczy, G. 1994, "Megjegyzesek a kora avarkori temetkezesi szokasokhoz (A flke

siros temetkezes)". In A k5kort61 a kzepkorig. Tanulmanyok Trogmayer Otto 60. szl
etesnapjara, ed G. Lrinczy, Szeged, 311-35.
)1-46.
1995, "Flkesirok a Szegvar-Oromdli kora avar kori temetbl. Kehany
Jiogica Am
megjegyzes a flkesiros temetkezesek vattozatair61, kronol6giajar61 es terleti elhel
yezkedeserl'; Mora Ferenc Muzeum Evknyve. Studia Archaeologica 1,399-416.
'eval Europe
Lrinczy, G., and Straub P. 2006, "Az avar kori padmalyos temetkezesrl. Szempontok a
karpat-medencei padmalyos temetkezesek ertekelesehd; Arrabona 44, 277-314.
Id ihre VerMontelius, O. 1903, Die lteren Kulturperioden im Orient und in Europa, 1. Stockholm.
Mller, R. 1996a, "Die Keszthely-Kultur': In Reitervlker aus dem Osten. Hunnen +
Awaren am
Awaren. Burgenlndische Landesausstellung 1996. Schlo Halbturn, 26. April-31.
'ogietransfer
Oktober 1996, ed F. Daim, K. Kaus, and P. Tomka, Eisenstadt, 265-74.
1996b, "ber die Herkunft und das Ethnikum der Keszthely-Kultur:' In Ethnische
~fassen". In
und kulturelle Verhltnisse an der mittleren Donau vom 6. bis zum 11. Jahrhundert.
Apostolski,
Symposium Nitra 6. bis 10. November 1994, ed D. Bialekova and J. Zabojnik, Bratislava,
75-82.
ngsbericht",
Renfrew, C. 1973, Before Civilization. The Radiocarbon Revolution and Prehistoric
Europe. London.
Shennan, S. 1990, Quantifying Archaeology. Edinburgh.
le des 6. bis
Siegmund, F. 2000, Alemannen und Franken. Ergnzungsbnde zum Reallexikon der
Germanischen Altertumskunde 23, 472p.
:on.
Stadler, P. 2005, Quantitative Studien zur Archologie der Awaren 1. Vienna.
.1.
Taylor R. E. 1987, Radiocarbon Dating. An Archaeological Perspective. Orlando.
ls. London/
ToCik, A. 1962, "Keramika so znackami na dne zo slovansko-avarskych pohrebisk na
juhozapadnom Slovensku'; Pamatky Archeologicke 53,347-80.
~eitlichen
T6th, E., and Horvath, A. 1992, Kunbabony. Das Grab eines Awarenkhagans. Kecskemet.
hcoming.
Varsik, V. 1992, "Byzantinische Grtelschnallen im mittleren und unteren Donauraum
ies in Early
im 6. und 7. Jahrhundert': Slovenskli Archeologia 40, no. L 77-108.
WallaceHadrill, J. M. (ed) 1960, The Fourth Book of the Chronicle of Fredegar with its
lfhunderts~
Continuations. London. Werner, J. 1950, "Slawische Bgelfibeln des 7. Jahrhunderts':
In Reinecke Festschrift zum 75. Geburtstag von Paul Reinecke am 25. September 1947,
kerdesben;'
eds G. Behrens and J.
Werner, Mainz, 150-72.
ischen Bev
- - . 1960, "Neues zur Frage der slawischen Bgelfibeln aus sdosteuropischen Ln
dern", Germania 37-38, 114-120.
Iked-Feket
---. 1962, Die Langobarden in Pannonien. Beitrge zur Kenntnis der langobardischen
Bodenfunde vor 568. Munich.
SdosteuroWild E. M. et al. 2000, "14C dating with the bomb peak: an application to forensic medi
eine': Nuclear Instruments and Methods. Section B: Beam Interactions with Materials
chungen,ed
and Atoms 172, 944-50.
Wood,1. 1994, "Fredegar's fables". In Historiographie im frhen Mittelalter, eds A. Scha
tnl A. Inns
rer and G. Scheibelreiter, Vienna, 359-66.
ia Bulgarica

)(}-11.

n polskieh':

Illustrations

'ugoslaviji,"

;ai Ertesit6

!Underts, ed

Figures
1. Image Database "Montelius", an example of the complex mode view: selected arti
facts from the qagan burial in Kunbabony (Hungary).
2. Image Database "Montelius': an example of the typological mode view: ceramic ware
with S-shaped, prick-like comb punch ornament.
3. A model for the creation of the Image Database "Montelius" on the basis of the pub
lished archaeological record.

"

82

PETER STADLER

4. A model of the possible uses of the Image Database "Montelius" for archaeological
studies.
5. Seriation by reciprocal averaging of over 4,000 Avar-age maJe burials.
6. Seriation by reciprocal averaging of the eigenvectors of over 4,000 Avar-age male
burials.
7. Seriation by correspondence analysis of over 4,000 Avar-age male burlals.
8. Seriation by correspondence analysis of Avar-age female burials.
9. Wiggle matching of radiocarbon dates with sequence dates from the seriation of
Avar-age burial assemblages.
10. Wiggle matching of radiocarbon dates with sequence dates from the seriation of
Avar-age burial assemblages.
11. Zoomed detail of the correspondence analysis scattergram of functional types of
artifacts from Avar-age burlai assemblages.
12. Distribution map of dress accessories with dentil decoration by ornamental motifs.
Data after Heinrich-Tamaska 2007.
13. Distribution map of dress accessories with dentil decoration by production tech
niques. Data after Heinrich-Tamaska 2007.
14. Distribution map of combs with teeth in a single-row deposited in graves.
15. Distribution map of combs with teeth in a double-row deposited in graves.
16. Distribution map of ceramic wares in the Carpathian Basin.
17. Distribution map of ceramic wares with prick-like comb punch ornament.
18. Distribution map of ceramic wares with potter's marks.
19. Plotting of the analysis of N next neighbors for all pottery features associated with
Middle and Late Avar burial assemblages.

NEW~

S~