Frontier Settlement and Economy in the Byzantine East Author(s): Michael Decker Source: Dumbarton Oaks Papers, Vol.

61 (2007), pp. 217-267 Published by: Dumbarton Oaks, Trustees for Harvard University Stable URL: http://www.jstor.org/stable/25472050 . Accessed: 28/06/2011 05:04
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at . http://www.jstor.org/action/showPublisher?publisherCode=doaks. . Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission. JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Dumbarton Oaks, Trustees for Harvard University is collaborating with JSTOR to digitize, preserve and extend access to Dumbarton Oaks Papers.

http://www.jstor.org

East Frontier and Settlement Economy in the Byzantine Michael Decker

1-^

his work

examines

on the seventh particular focus to to eleventh centuries.1 offers substantial promise Archaeology issues of settlement and land use, and these inform on the specific in turn lend considerable into the nature of frontier life and insight a conquests of the seventh society around the time of the firstMuslim and throughout medieval Byzantium. One of the major pre century of historians and archaeologists of Byzantium occupations today is economic condition of the empire around the time of the first the Muslim conquests, when the world of late antiquity crumbled and a new, medieval world ancient emerged from the wreckage of many structures. A clear of the land is vital to any of picture understanding the changes attending the arrival of the Byzantine dark ages and the in the ninth and tenth cen subsequent revival of Byzantine fortunes vast turies. Agricultural pursuits formed the livelihood of the major of the population and cities most often depended for their survival ity Material immediate agricultural landscapes. data that inform us of the kinds of farming and animal East also bear implications for Byzantine

tlement and agriculture from circa 300-1000, with

the archaeological evidence for set on the eastern frontiers of Byzantium

on their own

husbandry in the Byzantine at or lack of society large. They provide evidence for the presence nomadic elements, the potential for substantial settled populations, the persistence and character of sedentary communities, and of the network of social relations. The in further illumi archaeology nature and function of themiddle nating the Byzantine elite remains and there is no space to examine the largely untapped, although role of

paramount question of the expansion of aristocratic power, much of the picture rendered below is predicated on a belief that the eastern medieval aristocracy asserted great influence in the countryside. of Byzantium were not static. The limes over that the Roman centuries of Empire created piecemeal long experience in the East collapsed beyond retrieval in the seventh cen Persians and then under the
study will address other important sites, such as Dara, Apamea (Meskene), (Resafa), and Zenobia

The eastern borderlands

tury, first in the face of the Sasanian

In light of the theme of the Spring Symposium at which I first read the core of this paper, now considerably adapted, I have focused primarily on Cappadocia and the borders of Anatolia, namely the Tigris and upper Euphrates, although I stray into Syria i

and parts farther south for the sake of comparison. Due to space considerations, it is impossible to provide here a thorough review of the Roman-Byzantine limes; thus, while Antioch I have paid considerable attention to and the region of Chalcis, a future

excavated

(Afamia), Barbalissos Sergiopolis (Halabiyya).

/ /Amida

Dara

. / f

J .

( /

s \1

ResainaV /

\i

^ /
\
X.

(?

/T.Bati/

\ ( Vf /^'

T/Brak^/_ ^^
^ ( BirHaidarY
^ [

J J

^"~X> y

\
j-1

^%

X^^^^^KlTZonbil _ Tunainir Kh.HassanAga ^

AmostaeN <* ?1-Han \ Araban f y I

\ Lake Khatuniyeh \ cxtfl** \ )-Sl \

V T. Sufuq J
^"V T. el-Hamda ( \ Fadqami T. el-Ghail /

^~\^v

Ma

el-q.a'ara

\.

Fig. i The Tigris and Khabur on the eastern frontier. Reprinted from D. Kennedy, Rome's Desert Frontier from theAir (London, 1990), fig. 9b.

2l8

MICHAEL

DECKER

>-?.?'

Samosata

\ Edessa*

Zeugm?)

#Batnae

Carrhae# I ^

* Cyrrhus

\fW

V
Beroea Antiochtf ) Seleucia-*>^ \ C \ k Barbalissus?\ Callinicum / / *w-s^-J, T.el-Hajj*? J

/ Souriya ^^^\

* (
/ ? * "*

Laodicea 4-^
*^ ^

Resafa

\
C Apamea ' ?w?/vA \ ?* \ \#E1-Qdeyr \ *Nedwiyat \Q. %* el-Qdeyr

;
' Raphanaea

Hama*>^
^?T^ /v & v /

^Tayibeh >Aradus ( <*W ^>Sukneh

Mediterranean

/ Emesa / ^n / ( Palmyra^-*

3 ^

^/

Hleleh

?Tripolis

/ y

v /"*

^P /

/
f

^
V*"

Kh. el-Beida/ . TT . . _. Ummel-Amad ,. Qasr el-Heir el-Gharbi* n.0"* K.el-Haliabat * , ^## non.K, ^? / Danaba?

Harbaqa #Y?'

Qaryateyn J?

^
ZHdiopolis

\
?K.Aneybeh

#_/K.el-Qattar v /''Y.el-Basiri ^0' /

Giffa*

(Baalbek)
' ' ^.Dumeir, **

***

/K.el-Manqoura

.T"
^Damascus

el-Seyqal #Q. ..

o
0 700&w

>DeirSemali

_ .c

Fig. 2 The Orontes Valley and Syrian steppe along the eastern frontier. Reprinted from Kennedy, Desert Frontier, fig. 9a.

BYZANTINE

EAST

FRONTIER

ECONOMY

219

Koloneia , Fortresses/defended towns Towns abandoned during

this V period

_ JM^

. f,-s-^r'

. 1500 m elevation / _V^_?'?""*"?"\

S^ Thermalr Basilika ,_ V \
)

I-*&basteia
^X_. J J^^?'?
f/^^^v

(*s

^?J

Keltzine

-/

\_~_-^

^"^V^^^
{

l\ ^

Hisn Kamkr?"
(Kamacha)j
V/ /-?-'

C-?

/^

v-X)hozanon?

Nys^V
/

f^v^^Jj^ A^~~J

a^'^A^ Kaisareia
#

Nazianzos#

\fY . <$< -ir Kysistra $^

jfdf&i"^ ^y^#Tzamandos

/^SH^^

-?.
^^\ . .

(al-Qurrah) Tya|]fl ...

Koron.

RhoAnton V^A , J
J>*/
^\ Jtf j/

jT >_^\ \ J Ko4arT^Arabissos LJ \ HisanZibatrah V / J^^\ ^Afdn?S h-SJ V # \ ^-! ( } (Sozopetra) v r ~n ' \ wrasKisos [ / J UKiskisos-:i7 U^T^ #4KoUkousos ~?O/ i

S\ V' ?Osl \ \
yL_-

' ^-A #Turandah -C?S^^^ / (Taranton)

f W

J^"

His^Zi^^v^^imsat (Arsamosata)?
"\T (Charpete)

/ \J[ ) h\^iisn / i#Mal ^U QalVdhiyah ? \ ) (Melitene) / V C /.

f

DARB AL-HADATh,\\ ^^^ al-Hadath

(Ada\a)\/
Kaisun

CT-A I^ \V v( ( \ Amidak^^^V^ C

A.. TT '/ <$% Herakleia v PodandosV/ \ 7 (al-Badandun) fcc^/ SWT ^ ( (Kybistra) fc^tlKILIKipi^ -~-\T} \ larsus^s

Mar'as sfrv? A*

(Qermanikeia)

(D0l,Cl,C) .Harran \ I ^ ^^-^M^KWisah AdhlVah/(Mopsuestia) Qurus

~j J/ x."^,

f \ VSls!SW S& i / Y'AinZarbalTtAiiabarza)

/?-^^V-^

"23 X-^^umaysat

Vr*^^^^

JVO

] I

(Samosata)

,*>#,

S \ L

Ourfa (Edessa)

AL-JAZIRAH

(Kyrrhos)

,
^sSeleukia

\\
\. t

"^
v /

\S
/ /-^

V_\ ^^

(al-Iskanderun) SYRIA

Alexandretta

r (Hierapolis)

Manbidj^Xre ^V

^n.
-T^^?-s.

v-rx^-

Gulf of Tarsus

Halab(Beroea)

}>> .2takiyz (Antiocheia)

I Vx

100 km ^^^^

the huge areas of territory they controlled presented little obstacle to i conquest. The first earlier frontier (figs, and 2) stretched from the on the to Caucasus Syria, centered riparian districts of the Euphrates and Tigris Antioch, east of which and their tributaries, and extending as far as the region of the frontier melted into a blur of steppe and the fourth to the seventh centuries, the limits of the

We must speak of a permeable frontier onslaught of the Muslims. zone, with cities forming the primary nodes of political control, the vertebrae of the Byzantine backbone. Once these cities surrendered,

Fig. 3 The medieval frontier zone. Reprinted from J.Haldon, Byzantium in the Seventh Century (Cambridge, 1990), map 5.

desert. From early

an almost irresist presented ible tide of invasion, which broke itself only on the barriers of the Taurus and Anti-Taurus and thus the Romans found mountains, from the failure of the first: theMuslims themselves thrown back to the old frontier that had prevailed in the

occasions renegotiated, first by the of settled life to its greatest extent since the Bronze Age, expansion and second by the numerous episodes of warfare that interrupted the generally prevailing political calm. The second frontier was born

frontier were on numerous

220

MICHAEL

DECKER

^-O f / N-\

Batman DamA

^y^ ^^yv-jf

vfff ^^"^^ $(
\/

Diyarbakir?^ ^^--^_^J
?"N v

^\f
// J

IlisuDamiri r-( /

Ourfa ) Gaziantep Birecik Dam wk. *^ " Karkamis Dam V.-"' ..,

f^^^'

TURKEY
..

Mardm#
\ K^ "" IRAQ

k yf CizreDam^

......-

"" \^J^/

...-'* ^v

) SYRIA / ^N
" Dam projects
Country borders j ^ ^mm^^^^^m ^^

L^

,-

100 km

''

theMuslim conquest of the Byzantine days of Vespasian. Following a one foot in East, Cappadocia, giant upland territory that dipped the Euphrates and another in the salt marshes of the Toz Golii, once

Fig. 4 Upper Euphrates and Tigris region of the Byzantine frontier. Reprinted from G. Algaze et al., "The Tigris-Euphrates Project,"

resistance to an eastern enemy. again formed the heart of Roman 860 the Byzantine-Muslim Around frontier region (fig. 3) lay a line running between Sebasteia and Tephrike in the north and along the Cilician Gates and headwaters

River (modern of the Calycadnus in the south. By the eleventh century the offensive wars Goksu Nehri) of the Byzantines had advanced their territory about 150 km east of Trebizond with a prominent salient around Lake Van inVaspurakan, and thence southwest to the Orontes valley. This region is archaeo logically the wake

in part due to recent fieldwork conducted in interesting of Turkish hydraulic projects, but also because this rugged for landscape formed the interface between Byzantines and Muslims centuries and yet remained an inhabited where agrarian countryside and pastoral life continued. An investigation of the core elements of this rural activity comprises the present work.

Both

middle

centuries) and the the Tigris formed a Byzantine period (8th-nth centuries), barrier rarely encroached upon by Byzantine political control. It is not to examine the whole corpus of survey work from possible here period (4th~7th the Tigris-Euphrates corridor, and so I limit discussion to a number

The Tigris valley (fig. layat the limitsof theByzantineworld. 4)

in the Early Byzantine

Before

of illustrative projects that record data of interest for Byzantium. I proceed, it is important to note some of the promise and in the use of survey data in reconstructing problems archaeological

BYZANTINE

EAST

FRONTIER

ECONOMY

221

past environments and economies. Firstly, survey methods vary con siderably, and those that yield the best results generally incorporate local excavation, which Since many Tigris and Euphrates Likewise, more permits precise of the survey efforts undertaken were chronologies. in the valleys of the forerunners of dam projects, such excava ceramic

tion has not always been possible.

is always a In some cases, challenge. geomorphologi obscure sites; in others, later use may obscure earlier changes may is often difficult to Because knowledge of coarsewares occupation. to other, often obtain, they may be wrongly assigned early, occupa tion periods. levels based on numbers of sites is Interpretation of population as many also problematic, archaeologists have recognized, and thus the total occupied landscape area that belongs to a given period should site recognition cal also of any rigorous analysis. Geographical form the component is another concern: the data considered herein come from coverage areas of the or may not riparian districts, which may typify other ancient and medieval Levant. Despite these considerable challenges, as evidence in aims, continues to mount

survey-sampling strategies naturally have built-in biases must be considered in not all survey that handling the material, and ors' methods are entirely clear. No matter what themethod employed,

scope, and methods, a obtained broadly coherent picture of settlement and land exploita tion for the Byzantine and Early Islamic periods. Nor should we be surprised

from a variety of surveys that differ we can be confident that we have

variation is sometimes considerable, given the that regional and political conditions that prevailed from of geographical diversity the seventh to the eleventh centuries. These and other challenges will be met and overcome as surveys continue to grow in number and sophistication.2

The Tigris Valley

From

Nymphaios), theBohtan Su notably theBatman Su River (Byzantine

and his team worked along 1988 to 1990, Guillermo Algaze the course of the upper Tigris where it is joined by several tributaries,

preparationforportions of theGiineydogu Anadolu Projesi (GAP),

a scheme that encompasses hydroelectric and irrigation sizeable portion of eastern Turkey. After decades of construction and $32 billion invested, the face of the Tigris and Euphrates and many of their tributaries will never be the same. Unfortunately, Algaze's work the massive

Su. In addition Algaze and the Garzan explored (Byzantine Zirma), the region of the Tigris itself around the modern town of Cizre, all in

For references to the present state of and for the problems

survey in Mesopotamia

but overall utility of archaeological survey data, see T. Wilkinson, "Regional Approaches to Mesopotamian Archaeology: The Contribution ofArchaeological Surveys," Archaeological Research 8 (2000): Journal of 220-67.

has been too cursory to provide a clear sense of the archaeology of the affected area. The published material has some of the characteristics

222

MICHAEL

DECKER

of emergency work, with coverage necessarily determined In the late Roman/early

most quickly. that thefloodwatersfrom damswould affect

Byzantine period, the Batman

areas by the Su formed

an intensification of settle striking that the archaeological data show ment and land use, the textual evidence that portrays corroborating a the low river terrace lies flourishing region. Around Martyropolis

part of the hinterland ofMartyropolis (modern Silvan). It is thus

Su approaches the Tigris, the low terrace broadens and presents a wide band of easily exploited, rich alluvium. its fortification program, witnessed Via in the of writings on the remains of for in the sixth century and Prokopios drawing the Batman tifications of the medieval state made efforts to era, the Byzantine secure these farmlands and to protect the riparian crossing points that served as routes of trade and invasion.3 On the west bank of the Su at Semrah Tepe are found the remains of a fort that con trolled the crossing nearby at Malabadi. The finds from the imme area of Semrah included a coin of the emperor Phocas Tepe as well as a small amount of medieval (602-10), early Byzantine and and Islamic

a meters above thewaterline and is therefore only few easily exploited means of canals. It thus investment in by required minimal lifting or such as irrigating waterwheels buck gear (norid) counterweighted ets were available to the late antique inhabitants. As (shaduf) that

Batman diate

Byzantine

glazed wares, perhaps indicating that the Byzantines returned to the spot after the emir ofMartyropolis finally submitted to Byzantine authority around 976. Settlement apparently declined after theMuslim conquests and only later, during the period ofArtukid/Seljuk control (nth-i4th centuries), did settlement again no At that time the river to a conten peak. valley longer belonged tious frontier between two rival empires.4 Farther east

Su valley, along the frontier zone, in the Garzan sites yielded pottery of the late many Roman/early Byzantine period. Like the Batman Su region, the Garzan Su region witnessed heavy settlement during the fourth seventh centuries. Farther through downstream the Bohtan Su valley, the last of the surveyed tributar ies of the left bank of the traces of the brown Tigris, provided few and red wash ceramics that characterize late Roman/early Byzantine This is to the perhaps due region's situation: the Bohtan

occupation.

Su districtformed part of thehinterlandof Si'irt,

3 Prokopios, Buildings, 2.2.1-21;

part of the initially

Algaze, R. Breuninger, and J.Knudstad, "The Tigris-Euphrates Archaeological Reconnaissance Project: Final Report of the Birecik and Carchemish Anatolica Dam Survey Areas," 20 (1994): 1-96.

2.9.18-20; Dewing 4

2.4.3; 3.2.11-14; trans. H. B.

(Cambridge, Mass, 1971). G. Algaze, "ANew Frontier: First Results of the Tigris-Euphrates Reconnaissance Project,

241-81; G. Algaze, i988,"/iv~?S48,4 fo^): R. Breuninger, C. Lightfoot, and M. "The Tigris-Euphrates Rosenberg, Archaeological Reconnaissance Project: A Preliminary Report of the 1989-1990 Seasons," Anatolica 17 (1991): 172-240; G.

Archaeological

BYZANTINE

EAST

FRONTIER

ECONOMY

223

kingdom

of Armenia,

and then Persarmenia.

But the absence of late

antique pottery from the valley is somewhat perplexing, since there is at the fortress of at the conflu clear evidence of occupation (Jattepe ence of the and the Bohtan Su. Tigris In the main, survey around the plain of Cizre recorded Algaze's the result that the late Roman/early Byzantine period, in upper which is typically characterized Mesopotamia by small, dis a is persed settlement, underrepresented. The survey did record major late Roman/early fortress and associated settlement at Byzantine on thewest bank of the town Eski Hendek Tigris, 14 km north of the of Cizre. Eski Hendek was probably the late Roman fortress city of tell sites,with Bezabde, The site reminds us by Ammianus Marcellinus.5 wit of themilitarized qualities of the Tigris region, which frequently nessed Persian and Byzantine raids and invasion. Despite this, settle mentioned of the area under

ment does

over much seem to have been fairly lively control.6 Roman/Byzantine

Between the Rivers: TheKhabur-Ballkh Watershed

A an of archaeological projects have produced impressive in the river zones where settled com chronicle of human occupation number possessed two

munities

and steppe land meet. late antiquity, the valley of the Khabur During Byzantine urban hubs. In the north lay Theodosiopolis in the south, Circesium. These the

of the Khabur valley was length late antiquity, it is likely that another, as yet irrigated throughout in the middle reaches of the river unknown urban center developed This site is probably obscured by Islamic development.7 plain. the region around In the 1970s Deiderik examined Meijer course of the in Syria (ancient Qamishli Djaghdjagh along the a of the Khabur. The surveyed regions River, tributary Mygdonius) hinterland ofNisibis, and after 363 portions of the covered area prob in Byzantine hands. After 507 this segment of territory was exten from Anastasiopolis/Dara. Meijer's work period,
Rerum

Resaina), the length of the river. Because

(Ras el-Ayn/ two cities bracket, as it were,

of the Djaghdjagh valley and the adjacentplains formedpart of the

ably remained was governed sive; more than 300 sites were to the Early Byzantine
5 Ammianus Marcellinus,

recorded. By my tally, 53 sites belong with 18 categorized broadly compared

"The Site of Roman Bezabde," ed., Armies and Frontiers in S. Mitchell,

gestarum 6

libri qui super sunt, 20.7.1; 20.11.6; Seyfarth (Leipzig, 1978). Su/Cizre)

inRoman and

21.13.1, ed.W.

(Bohtan Su/Garzan

Byzantine Anatolia (Oxford, 1983), 189-204. 7 J. Lauffray andW. Van Liere, "Nouvelle prospection archeologique dans la Haute-Jezireh (1954): 129-48. syrienne," A A rchArSyr 4-5

"The Tigris-Euphrates Archaeological Reconnaissance Project," 187-99; (Eski Hendek-Fenik/Bezabde) C. Lightfoot,

224

MICHAEL

DECKER

180?1 160-1-

i4oj ,20H-1loo-80--

? I i6n-,
-

14-.
12-i 10?I

.lll..ll
oPv

i?I
25-t-1

-111

ll

30?1-1 20

lllllllH

JiTil-bMiIIII

into "Hellenistic-Parthian-Roman" and 93 of the "Islamic byMeijer To the west of this zone, the Leilan Survey (fig. 6) Period" (fig. 5). found that evidence for the Byzantine period was triple that of the Roman period the survey sample and continued in the Islamic period. so should not be is small and overemphasized.8 instead to increase But

Fig. 5 Northeast Fig. 6 Leilan

Syria survey

survey

Fig. 7 Tell Brak survey Fig. 8 Tell Beydar survey. After T. "Tell Beydar Survey, 1998-99 Report," http://oi.uchicago.edu/ research/pubs/ar/98-99/beydar.html.

Wilkinson, Annual

and

The Tell Brak survey (fig.7) did not periodize earlyByzantine

Islamic material, to these later settlement referring can "First Millennium." One rubric of

Like the study around Tell Brak, theTell Beydar survey (fig. 8)
data produce a similar image of settlement a combination some visible variations. Using over two seasons, surface reconnaissance survey recorded 82 sites. After in the midst of the contested lay stability, though with of remote sensing and

phases under the general therefore gain only an impressionistic view of broad continuity in era settlement from the Hellenistic period. through the Muslim

Beydar scape

Tell Tony Wilkinson's the treaty of 363, this land area between the fortress

D. Meijer, A Survey in Northeastern 1986).

Syria (Istanbul,

BYZANTINE

EAST

FRONTIER

ECONOMY

225

35 ?I-1

40-^^H-^^=-

_ !_H^^B_

I I

I I

of strong Hellenistic, Parthian, and the latter despite the posi Byzantine-Sasanian period habitation, on the boundaries tion of this of the empires between landscape Nisibis and Byzantine territory.9 in the pattern of human Changes exploitation of the landscape did occur, however. From the Hellenistic through Byzantine/ Sasanian

city of Theodosiopolis (Ras el-Ayn) and Nisibis (Nusaybin). The

surveyors recovered evidence

Fig. 9 Khabur survey (Upper Khabur). After B. Lyonnet, "Settlement Pattern in the Upper Khabur." Fig. 10 Middle Khabur Valley

site numbers declined and sites periods, aggregate larger to smaller settlements. This habitation and gave way dispersed lack of large central urban centers parallel situa the above-noted

were found, but none of these were sited on tells. The ("Sasanian") area continued to be fairly densely settled during the Early Islamic but not as actively as in previous period, ages.10 From 1974 to 1978, Bertille Lyonnet collected data from the Khabur, focused where

tion that prevailed along the Tigris. The material from Tell Beydar to include the need for surveys in Mesopotamia underscores low-level sites in their thirteen late antique sites methodologies:

from the Khabur, was led by archaeolo we in earlier interested gists periods, but the evidence glean from them is nonetheless valuable. Jean-Yves Monchambert's work most focused on the middle Khabur period, the Hellenistic/Roman to the Islamic period

found and 14 sitesofUmayyad-Abbasid date (fig. This fieldwork, 9).

on tells, like most

29 Byzantine

sites (14 of these are in question)

were

to valley, where 30 sites belonged 11 to the and 15 (early) Byzantine,

In the lower stretches of the river (fig. 10).

J. Eidem and D. Warburton,

"In the

Land ofNagat: A Survey around Tell Brak," Iraq 58 (1996): 51-64.

"Archaeological Subartu

(Tell Beydar) T. Wilkinson, Survey of the Tell Beydar Region, Syria, 1997: A Preliminary Report," 6 (2001): 1-37;Wilkinson, "Tell Beydar Survey: 1998-99 Annual /98-99/beydar.html. Report,"

io

http://oi.uchicago.edu/research/pubs/ar

226

MICHAEL

DECKER

in the territory of Circesium, data from surveys once more

preliminary

140-i-1

suggest that, in the Hellenistic-Parthian peri beginning Site ods, settlement was more dispersed. size also decreased the Hellenistic, spatial distribution environment. This ists fanning in hamlets out over when periods predating tells dominated the the human-built the agricultur length of the

of

trend, with

valley

persisted with through the Roman,

along and villages on the plain, apparently little interruption

west of the Khabur is the Balikh Excavation Euphrates (fig. 2). to and and survey work in this region has focused on periods prior after Byzantine site the latter notably on the Abbasid occupation, ofMadinat to Karin sites in the valley. According cen valley survey, the period from the first a inten tury BCE to thirteenth century CE marked long phase of was the sive settlement. The densest period of occupation Early al-Far and other Balikh Islamic Bartl's era, to which

and early Byzantine, was appar early Islamic periods. The latter ently particularly well settled (fig. ii).11 next The of the major tributary

Fig. ii Lower Khabur Valley

Islamic

large urban of occupation

new town of Madinat al-Far, a belongs the center (more than 100 ha in extent). The precise range at the site is uncertain. While the Umayyad foundation the excavators of Hisn Maslama identify known

the town with

are Abbasid from literary sources, the stratified finds as published in date.12 Future work will almost confirm the earlier certainly in fact, if the of Madinat al-Far. It would be unsurprising, origin site In other predates the Islamic conquests regions of altogether. the steppe lands in the transitional and the steppe, other sites bear zone between the settled lands to Byzantine striking testimony

Balikh valley sites, Tell Sheikh period vitality,including important

ii (Upper Khabur) B. Lyonnet, "Settlement Pattern in theUpper Khabur to the (N.E. Syria), from theAchaemenids 'Abbasid Period: Methods and Preliminary Results from a Survey," in Continuity and Northern Mesopotamia from the Change in Hellenistic to theEarly Islamic Period, ed. K. Bartl and S. R. Hauser (Berlin, 1996), 349-61; (Middle Khabur) J.-Y. "Le Moyen Khabour: 37; (Lower Khabur) Frey, H. Kiihne, Reconstruction J. Ergenzinger, W. and H. Kuschner, "The W. Rollig and H. Kiihne, "Lower Khabur: Second Preliminary Report on a Survey in i977,"^4^4rc/?^r,S)/r33 (1983): 187-99; most recently (unseen): H. Kiihn, Magdalu/Magdala: Tall Seh Hamad von derpostassyrischen Zeit bis zur romischen Kaiserzeit (Berlin, 2005). 12 K. (Madinat al-Far/Hisn Maslama) Bartl, "The Balih Valley, Northern Syria, during the Islamic Period: Remarks Concerning the Historical Topography," Berytus 41 (1993-1994): 29-38.

of Environment, Irrigation and Development of Settlement on the jabur inNorth-East Syria," in Conceptual Issues inEnvironmental Archaeology, ed.]. Bintliff, and E. Grant (Edinburgh,

D. Davidson, Kiihne,

1988), 122, fig.8; see also W. Rollig and H. "The Lower Khabur: A Preliminary

Monchambert,

Prospection preliminaire a la construction d'un barrage," AArchArSyr 33 (1983): 233

by the Tiibinger Atlas des Vorderen Orients in 1975,"AArchArSyr 32 (1977/1978): 115-40;

Report on a Survey Conducted

BYZANTINE

EAST

FRONTIER

ECONOMY

227

Hasan,

Khabur thatpersisted through foundation (Dur Katlimmu) on the

the Byzantine period. Settlements elsewhere in the steppe zone in in the steppe lands the desert or transitional areas, like Androna, south of Chalcis, Bartl's or Euaria, near Emesa, the Balikh survey work grew considerably.13 recorded 80 siteswith

40 km north of Raqqa,

or Tell

Sheikh-Hamad,

an

Assyrian

sites are labeled Byzantine Period (4th~7th centuries) and another 14 as In that the early Islamic "potentially" late Roman/early Byzantine. in many of the surveys discussed is fairly poorly represented period

traces along in the of 10th of occupation early Islamic period (mid 8th-beginning of which were securely dated to this time and 25 are prob century), 55 sites are securely dated to the late Roman/early able. Twenty-three

fromthe Tigris and the landsbetween itand theEuphrates, thefinds

the Balikh present an interesting anomaly. along There are several possible reasons for these findings. The first, and

most is that some material classified in other surveys likely, perhaps era. A second to the as possibility Early Islamic "Byzantine" belongs is sitemigration, with people who lived in the more exposed frontier farther inland regions of the upper Tigris and Euphrates moving or as part of an to the Balikh, organized perhaps either voluntarily of relocation initiated by the caliphal authorities. Such movements the conquered populations were not uncommon within theUmayyad or in the state. Settlements were Euphrates newly founded expanded The Umayyads and earlyAbbasids paid particular attention to valley. resided at and further developed confluence. Al Mansur Balikh-Euphrates (d. 743) Rusafa, not far from the constructed al

middle Euphrates,becauseHisham b.Abd al-Malik theregionof the

(d. 775)

site of Irrigation early Byzantine period Nikephorion/Callinicum.14 attended the development works and other agrarian development the precise period of peak population, of these cities. Whatever
13 C.-P. Haase, "Madinat al-Far: The of an Early Mango, "Excavations and Survey at

model ofBaghdad to replacetheoldHellenistic Rafiqa (Raqqa) on the

Nordsyrien

(Berlin, 1994); (population

Regional Late Antique Tradition in Bartl and Hauser, Islamic Foundation," "Madinat

Androna, Syria: The Oxford Team 1999," DOP 56 (2002): 307-15 (after I completed this work, the publication of the Harran Survey work ofN. Yardimci came tomy attention, but Iwas unable to take these data into account [N. Yardimci, Harran Ovasi Yuzey Arastirmasi/Archaeological Harran Plain, 2 vols. [Istanbul, Survey in the 2004]); M. M. Mango, "Excavations and Survey at Androna, Syria: The Oxford Team 2000," DOP 14 57 (2003): 293-97. "Prospeccion en el (Siria). Informe provisional," M. Cordoba, J.

transfers) Al Baladhuri, The Origins of the Islamic State, ed. and trans. P. Hitti (New York, 1916), 253; (Raqqa/Nikephorion/ Callinicum) M. Meinecke, "Raqqa on the Euphrates, Recent Excavations Residence Near East of Harun er-Rashid," inAntiquity: German Contributions to theArchaeology Syria, Lebanon, (Amman, S. Kerner al-Khalaf at the in The

Continuity and Change, 29-38; C.-P. Haase, al-Far?First Archaeological Soundings at the Site and theHistory of an Umayyad Domain inAbbasid Times,"

inBilad al-Sham during theAbbasid Period, Proceedings of theFifth International on the History ofBilad al-Sham, Conference ed. M. A. al-Bakhit and R. Schick (Amman, 1991), 1: 206-55; Valley (Balikh) K. Bartl, "Balih of the Late Survey?Settlements and Islamic Period,"

Palestine,

Jordan, of and Egypt, ed. 1990), 2:17-32; M.

and K. Kohlmeyer, "Unter suchungen zu ar-Raqqa-Nikephorion/ DM 2 (1985): 133-62.

Roman/Early Byzantine in Bartl and Hauser, Continuity and Change, 333-48;

valle rio BallhAula Orientalis Fruhislamische

Callinicum,"

(Androna) M. M.

6 (1988): 149-88; K. Bartl, Besiedlung imBalih-Tal/

228

MICHAEL

DECKER

45
40-g-

-i-1
120-.

^^^^^^^^^H

^^^^^^H

5S" ^

J?

i^>

-^

"-^^^c?

Total Sites Visited

"Roman Sites"

we have evidence zone

for considerable

during the sixth-ninth settlements, the creation of new fortified sites, palace attendant agricultural installations.

in this part of the frontier change with the foundation of new centuries, building, and

Fig. 12 Sajur survey Fig. 13 Jabbul survey

TheEuphrates Valley

A number of surveys along the Euphrates offer data that permit scope for comparison, both in the Euphrates valley itself, and with the north Jazira material. The Sajur Survey (fig. 12),which canvassed much of of early the territory of the city of Hierapolis (Membij), the capital Byzantine

Byzantine period thandid anyother.The Jabbul Survey (fig.13)also

offers only rough-grained data, but these again are evocative: of the remains. Over sites surveyed, more than half bear traces of Roman of the Levant, there was early Roman down therefore indicates that this portion of the province of Euphratensis was well region populated until the medieval period.15

Euphratensis,

recorded more

settlements from the Roman

most

strong settlement continuity from the to the early Byzantine period. The Jabbul Survey

and his team explored a 60 km stretch of the river valley Algaze area contained the late to Halfeti.16 The from Carchemish surveyed cities of and Samosata. Roman-/early Byzantine-period Zeugma is now well known for the richness of the mosaics the emer Zeugma there as well as for its tragic loss. The finds gency work uncovered
15 (Sajur Survey) P. Sanlaville, ed.,Holocene North Syria (Oxford, 1985); Settlement in (Jabbul) R. Maxwell-Hyslop, J.du Plat-Taylor, M. V. Seton-Williams, and J.D. Waechter, "An Archaeological 1939 "PEQ,74 Survey of the Plains of Jabbul, (1942-43): 8-40. \6 G. Algaze, R. Breuninger, and "The Tigris-Euphrates

J.Knudstad,

Reconnaissance Project: Archaeological Final Report of the Birecik and Carchemish Dam 1-96. Survey Areas," Anatolica 20 (1994):

BYZANTINE

EAST

FRONTIER

ECONOMY

229

so far suggest a Roman urban center that persisted in its flourishing to the end of the prosperity early Byzantine period. Recon the Around Zeugma Tigris-Euphrates Archaeological naissance Project noted "a substantial peak in the population of the Birecik-Carchemish Among found a number of dispersed settlements, perhaps farmsteads or villas. across the The spread of landscape generally indicates outlying farms less fertile and intensive land use, as farmers colonized both security area" (fig. 14) during the early Byzantine period. were traces of this late the antique population expansion

to do so. The over areas of the city landscape and felt safe enough all picture is one inwhich both town and country flourished in late a in theUmayyad antiquity, with precipitous drop population during There are several explanations for the lack of evidence for the period. earliest phase of Islamic control. As mentioned above, there appears

in the to be broad continuity inmany of the ceramics manufactured sixth and seventh centuries. Thus wares classified late Roman/early

the Byzantine may obscure early Islamic occupation. Alternatively, that lack of recognizable Umayyad pottery may reflect depopulation to came about due to Byzantine territory, deportation by emigration or death inwarfare and its attendant miseries. the Arab authorities, Whatever the case, Zeugma does not feature prominently in the tex it seems that the city tual sources of theMuslim probable period, and the waned after conquests.17 Three

other projects along the upper Euphrates produce results settlement. Survey that validate the picture of dense early Byzantine found few Roman remains, of the Keban Reservoir region (fig. 15) the middle period) was well rep while Byzantine material (including is the evidence from Kurban Hoyiik resented.18 More (fig. revealing Byzantine period had the greatest 16), where the late Roman/early era. The settlement extent of any findings from the Titrif Regional likewise reveal heavy settlement: late Roman/early Survey (fig. 17) it is sites far exceed those of any other period, while Byzantine
17 G. Parthey, ed., Hieroclis Synecdemus etNotitiae Graecae Episcopatuum (Berlin, 1866), 41; H. Gelzer, ed., Georgii Cyprii et al., "Mission C. Abadie-Reynal de Zeugma. Rapport sur la archeologique campagne de prospection 1995,"AnatAnt 4 E. Bucak, (1996): 311-24; C. Abadie-Reynal, Vallee E. Bulgan et al., "Zeugma-Moyenne de l'Euphrate, rapport preliminaire de la campagne de fouilles de 1999," AnatAnt 8 et al., (2000): 279-338; C. Abadie-Reynal des campagnes de "Rapport preliminaire fouilles de 2000," AnatAnt 9 (2001): 243 305; D. Kennedy, ed., The Twin Towns of on theEuphrates: Rescue Work and

Historical

Studies

(Portsmouth, R.I.,

1998);

R. Early and J.H. Humphrey, eds., Zeugma: Interim Reports, Rescue Excavations Institute): Inscription (Packard Humanities Mars, House ofAntiochus I, Bronze Statue of and Mosaic of the Synosai, and Recent Work on theRoman Army at Zeugma (Portsmouth, R.I., 2003). R. Whallon, An Archaeological Survey East-Central oftheKeban Reservoir Area of 18 Turkey (Ann Arbor, Mich., 1979).

Descriptio orbis Romani (Leipzig, 1890), 871-75; A. H. M.Jones, Cities of theEastern Roman Provinces, 2nd ed. (Oxford, 1998), "Sumaysat," EP 9: 871b; S. Redford, The Archaeology of theFrontier 532; C. P. Haase, Medieval of the On Zeugma, Near East: Excavations at Gritille, Turkey (Philadelphia, 1998), 5-8; am see J. Wagner, Seleukia (Wiesbaden, 1976);

EuphratI Zeugma

Zeugma

230

MICHAEL

DECKER

250?,-.

_ Fig. 14 Birecik Carchemish Survey region. 230_ After Algaze et al., "Tigris-Euphrates 220- Project," fig. 14. Reservoir

200180 160-mm? 150-? 140-? 130-_ 120110-? 100-35-1 40

& 190-

Fie. 15 Keban

Fig. 16 KurbanHoyiik _ Fig. 17 Titris regional survey

60-

I? 701 -

90-30-T-M-| 20-??

25

j.

Hi

?iii^i?=i=

35 25?1-1 30

?1-1

20-1-

2S

&y ^ ^ ,-c ,-c cT fc ,qd ?jfffffff^fiff^-3 ftr ? C ^ ?> P ^"ll'iTi'l'fl1 ?^ ? * >8 ^ 8 ^ *" # ? ? *2fZ * ^

^o ^

^ 'il *8

BYZANTINE

EAST

FRONTIER

ECONOMY

23I

perhaps

survey in southeastern the wealth and variety Turkey intensive systematic of sites requiring investigation. Unfortunately, in many areas, and the survey was never undertaken picture thus Results from Umit from Kemaliye Serdaroglu's broad to Keban underscore remains qualitative. Nevertheless n represented. Approximately ceramics, with substantial sites known we find the Byzantine period well of 58 sites yielded early Byzantine near Malatya (Melitene) and

earlyIslamic (Abbasid) periodwere found.19

no certain telling that

traces of habitation

belonging

to the

(Samosata).20 in the midst of the fertile Karababa Samosata basin, where lay water ren a pleasant climate, long growing season, and abundant dered the hinterland fertile to the extreme. Alluvium deposited by further sustained frequent springtime flooding of the Euphrates From its earliest history, then, the crops of this stretch of the valley. Samosata was probably an agricultural center rather than the locus of or extensive animal husbandry was ranching activity, though herding center The city's importance as a communication always prominent. was marked; and it commanded the only crossing between Melitene and (Hierapolis) Manbij across the river to Edessa the only convenient route from Anatolia thus (Urfa), 50 km to the south. Samosata

Samsat

its agri communication with Edessa, where it marketed enjoyed easy to theMuslims in 639/40, but in 958 cultural surplus. The city fell it. Samosata became part of the theme of John Tzimiskes reconquered center and a Byzantine administrative "the poleis on the Euphrates" of control in and consolidation and staging post for the expansion to capture in 1031 forces marched from Samosata imperial Edessa. After the battle of Mantzikert (1071), the center fell under control and then passed into Crusader hands, never to return Seljuk at both cities was to the practiced Byzantines. Irrigated agriculture Late Antiquity, probably with the assistance of water throughout because the depth of the banks of the Euphrates lifting machines, canal building in all save the lowest river terrace areas. challenges the East: at Samosata, and Islamic geographers extensive Roman noted that the agri

water-screw medieval isknown

a Roman period, the In the Legio III Gallica engaged inbuilding

The and irrigation. cultureof the landscapedepended on both rainfall

fate of Samosata's to have run more than 40 km along is unknown. The omission aqueduct theKahta system,which River and the
19 G. Algaze, A. Misir, and T. J.

indicate geographers may the aqueduct functioned in some disuse after the conquest. However, as late as the sixth century, when an inscription recorded the capacity a stretch of it. Samosata remained a frontier city until the repair of Turks overran the Byzantine East in the eleventh century. The loss of

right duct by theMuslim

bank of the Euphrates,

of the aque that ithad fallen into

Wilkinson, of California Anatolica

"?anliurfa Museum/University Excavations and Surveys at

Titrish Hoyuk, 1991:A Preliminary Report," 18 (1992): 33-60.

20 U. Serdaroglu, Asagi Firat Havzasinda Arastirmalar: 197s Surveys in theLower Basin (Ankara, 1977). Euphrates

232

MICHAEL

DECKER

more the aqueduct suggests a medieval city much modest than its late antique predecessor.21 now sub The remains of ancient Samosata, once covered 80 ha. There were both upper merged, and lower cities. The upper city mound, which the oldest settlement at undoubtedly comprised the site, once rose some 45 m above the river. The an area of 3 ha, was par which occupied excavated in 1964-70. The latest occupation tially to at Samosata was apparently the Seljuk period, an abundance ceram which belonged of medieval mound, I). Traces of the (Level III) were noted early Byzantine occupation exca along the perimeter of the citywalls. Further at Samosata in 1978-87; vations were conducted (Level ics and structural remains

80-1-1 70-? 60-? 50-? -j -

0?*

I*

I"

J ^

<" ^

I-1-1

- 4? ^

" I

? ir

I"

circuit wall. Despite the these investigated the 5 km long Roman of the site and the years of work conducted there, archae importance remain ological data from Samosata meager.22 and Savas Harmankaya's Mehmet survey Ozdogan and its surroundings noted enclosed portion of Samosata dated to the Classical period and 26 from the "Post-Classical" and stated that most but more explored of the

Fig. 18 Adiyaman

survey

12 sites

pottery at 73 sites (compared with 74 yielding Roman wares?fig. was at Tille of which the Adiyaman 18). Excavation Survey Hoyiik, a an extension, check on the survey pottery. provided chronological itmay be inferred on site numbers alone that the population While did not suffer a dramatic unknown increase or decrease from one period did not increase markedly settlement the dispersed to another), Roman over the Hellenistic characteristic of

the plains south and east of the modern town of Adiyaman, that in antiquity formed part of the hinterland of Samosata. territory The preliminary finds of this survey discovered medieval period

period of the latter sites were Byzantine-early Islamic, refinement is not possible. The Adiyaman survey further

(total occupied area is settlement apparently

period. Interestingly, the Roman and late

Roman

contrasts with that of the medieval period, when period new centers arose, often around fortified sites. One might recall that
2i ODB (Samosata) A. Kazhdan, "Samosata," Technical 1978-1979; 284-89. 22 M. Ozdogan, Lower Euphrates Basin 1977 Survey (Istanbul, 1977), 117;T. Goell, "Samosata Archeological Excavations, Turkey, 1967,"National Geographic Society University, Asagi Firat Projesi (^alismalari: Lower Euphrates (Ankara, 1987), Research Reports 1967 (1974): 83-109; U. Serdaroglu, Asagi Firat Havzasinda Arastirmalar: 197$ Surveys on theLower Euphrates Basin (Ankara, 1977); A. Tirpan, "The Roman Walls of Samosata," in The Eastern Frontier of theRoman Empire, ed. D. French and C. Lightfoot (Oxford, 1989), 2: 519-36.

3:1836; E. Honigmann, Die Ostgrenze des byzantinischen Reiches von 363 bis 1071 (Brussels, 1935); G. LeStrange, Palestine Moslems: From A.D. 6$o to i$oo under the

Project 1978-1979 Activities

(Beirut, 1965), 535; (aqueduct) U. Izmirligil, "The Samosata (Samsat) Aqueduct Investigations, 1978-1979," in Middle East

BYZANTINE

EAST

FRONTIER

ECONOMY

233

certain parts of the sites Byzantine Tigris Valley, where Roman/Early grouped around fortified points and latermedieval this pattern settlement was was scattered. One can based region on the limited evidence, less threatened only speculate, that the Samosata

is at variance with

18_i-1 16-1 14-' 12 10 8--

by Algaze some distance secure in the countryside dwelling from major military points.23 10 km Gritille on the Euphrates, upstream of shared with the latter the space of the Samosata, fertileKarababa Byzantines returned toMesopotamia

than the regions explored and that the late antique inhabitants felt

basin.24 Gritille was fortified by the during the eleventh century when they for the last time: the sub

sequent phases belong to the twelfth and early thir teenth centuries, with abandonment by the middle

^ # i ? - $? i ^ * i * t | i i ^#?g 2a? $ ? i <2f & ~?

Fig. 19 Gritille survey

i? *

-J? ^

$" ^

c?

.^r ?

i? &

of the latter century. Landscape survey that covered a 43 km2 area around Gritille on the west bank of the Euphrates revealed that the late Roman/early Byzantine period and the late medieval period were sites (fig. 19) that a total area of 10.73 na> to the Early Byzantine period covered belonged centuries yielded while themedieval phase of the eleventh-thirteenth same area (10.72 ha). two Abbasid seven sites Only covering the only both characterized by peak settlement. Seventeen siteswere known. We cannot completely rule out the possibility period that the Umayyad period is invisible to us, and that settlement in the

Karababa However,

on the fortified points of Gritille and Samosata and pation depended of the Seljuk state in the face of the quickly waned with the collapse invasions. Mongol

to the land worked over the countryside, closer period remains, including the by the farmers of the district. Medievalmore unsettled conditions: occu brief Byzantine reoccupation, suggest letswere distributed

as noted farther north, once attests to the same dispersed population secure conditions in which farmsteads and ham testament to again

basin persisted at a high level after the end of Byzantine rule. the late Roman/early Byzantine occupation around Gritille

Antioch andNorth Syria in the Early Byzantine Period

Antioch was

the key to northern Syria and Cilicia. It held a particularly important place within the strategic geography of the early Byzantine was much contested. Persian armies attacked Antioch in period and 529, and the city was sacked in 540. The Sasanians again besieged suc the city in 573 and seized control in 609/10. In 636/37, Antioch to theMuslims under their control until 969. cumbed and remained second period of Byzantine occupation lasted until 1078.

S. R. Blaylock, D. H. French, and 23 G. D. Summers, "The Adiyaman Survey: An Interim Report," AnatSt 40 (1990): 81-135. 24 Redford, "Archaeology" (above, n. 17).

The

234

MICHAEL

DECKER

to the sea via theMediterranean ports against Persia with its outlet of Seleukia Pieria and Antioch. the fourth-seventh centuries, During Antioch remained a vital civil administration center, the capital of

theEast, and thehead of one of the supplyroutesof theeasternfront

In late antiquity,

the city was

the seat of the

magister militum

of

where many of themore than seven hundred early Byzantine villages have been explored and so vividly detailed by the work of Tchalenko, Tate, and others. The produce from the Limestone Massif, mainly olive oil and wine, as Antioch, other served the needs of both the

a terminus on a Syria Prima, major routes of trade, and market for a the agricultural produce of In the midst of flourishing hinterland. this territory lay the low chain of limestone hills (Limestone Massif)

and Chalcis, and Apamea, overseas markets. was upon Surplus production predicated investment within the and heavy countryside, political demographic or overseas markets where the stability growth, and the existence of an outlet. agricultural surpluses found It is difficult and its Antioch to determine Antioch neighboring was not the same as its late city antique predeces which had a population intra muros of about 150,000, and sor, prob same number in the suburbs. This is to say ably the living nothing of the rural population who made their home in the large plain of the considerable inter surrounding hills. Despite est shown in the Limestone Massif, scientific survey and widespread excavation have been the Limestone mid-sixth slow to advance flourished there. Tchalenko the Persian believed that Massif until invasions of the and the the hinterland of precisely when of Apamea declined. Certainly city

regional towns, such also and Constantinople

medieval

Antioch

in the century, and then declined markedly early-seventh Sasanian occupation which, he cut the Antiochene century argued, off from its natural market hinterland outlets around the Tate's research showed that settlement in peaked the later fifth century and the mid-sixth century had decreased by More Foss has significantly. recently, suggested that the combina tion of invasion and plague caused substantial attrition among the in the sixth century and a dwindling of population until

Mediterranean.

population

Villages antiques de la Syrie du Nord: Le massif du Belus a I'epoque romaine, 3 vols. (Paris, 1953-1958); G. Tate, 25 Les campagnes de la Syrie du Nord du He au Vile siecle: un exemple d 'expansion demo graphique et economique dans les campagnes a la fin de I'antiquite (Paris, 1992); C. Foss, "Syria inTransition, A.D. 550-750: An

G. Tchalenko,

51 (1997): Archaeological Approach," DOP 189-269; M. Decker, "Food for an Empire: Wine and Oil Production inNorth Syria," in Economy and Exchange Mediterranean in theEast during Late Antiquity: Proceedings ofa Conference at Somerville College, Oxford, 29th May, 1999, ed. S. Kingsley and M. Decker (Oxford, 2001), 69-86.

BYZANTINE

EAST

FRONTIER

ECONOMY

235

was in the 630s.26 the coming of the Muslims Trombley, however, more in the Antiochene, optimistic in his assessment of occupation and using epigraphic data he concluded that the region maintained a substantial monks only population into the eleventh and agriculturalists Syriac-speaking century, with Arabization taking place of

in the twelfth.27

It seems likely, however, that when the Byzantines returned to a in the tenth century, at the Antioch they entered minor city fringe of the caliphate, not a of the booming provincial capital. Discussion in the fate of Antioch of cities and settlement elsewhere is neces light settlement in the Levant, since sary in any consideration of medieval recent work has at the heart of the discussion of urban placed the city transformation. Medieval Antioch still contained at least some of the monumental lation was buildings ascribed to it in late antiquity, but its popu close to its late antique peak: a tenth

century, been 40 m wide, until the city that had, in the mid-sixth was covered classical grid structures.28 Jean Lassus dated by medieval this activity to after 636,29 but recently Hugh Kennedy andWolfgang this interpretation. Relying primarily on Liebeschuetz challenged and Apamea evidence from Gerasa, Pella, Antioch, they argue that, at the end of in the Byzantine rule, illicit building increasingly closed most cities. The fact that the cities of Byzantine formerly open grids of no resistance to the Muslim invaders is also Syria-Palestine offered raised in the debate on the transformation of urban space; both phe products of ineffective government. Kennedy the swift conquest as reflective of long-standing demographic state at the time that gripped the Byzantine and economic malaise to this line of argument, cities were no of the conquests. According centers of trade or artisan activity on any scale, nor were they longer home to a prosperous or numerous elite who engaged in trade. Their is therefore in the face of theMuslims unsurprising.30 collapse viewed
26 (Population ofAntioch) G. Downey, A History of Antioch in Syria: From Seleucus to theArab Conquest (Princeton, 1961); E. Will, "Antioche sur l'Oronte: Metropole de l'Asie," Syria 74 (1997): 99-114; C. Foss, "Syria inTransition," 202. 27 F. Trombley, Cultural Transition Antioch, (2004): "Demographic in the Territorium and of 5 For an interesting analysis of the relics of the plan of Antioch in today's town, see 28 P. Pinon, "Permanences et transformations apres 5 (2004): dans la topographie d'Antioche l'Antiquite," TOPOISupplement 191-219. 29 J. Lassus, Antioch on-the-Orontes, vol. 5, (Princeton, 1972), Les portiques dAntioche 136-37,149-50. 30 (Decline ofAntioch and Apamea andW. and

certainly nothing to is of fifty thousand century population seventy-five thousand a reasonable The fall in population was not the only guess. probably on the great porticoes of the change. Buildings gradually encroached

nomena

are viewed

as

their hinterlands) H. Kennedy Liebeschuetz, Northern

"Antioch and theVillages

of

Syria in the Fifth and Sixth Centuries A.D.: Trends and Problems," Studies 32 (1989):

Nottingham Mediaeval 65-90; Tchalenko,

6th-8th c," TOP OI Supplement 341-62.

Villages antiques, 68-75 (above, n.25); C. Foss, "The Near Eastern Countryside in Late Antiquity: A Review Article," in J.H. Humphrey, ed., The Roman and Byzantine Near East (Portsmouth, R.I., 1995), 213-23.

236

MICHAEL

DECKER

10 ?

ii ?? ,

I?I | |

I?

i OS O r<^ vf

I?

I N kf

i ^

I p^P

iii

I?

rr\\ cO |

|\ rr*

iTM^I xf ^?f

?'N \\r\\ \\T\

t. i?iii \0 r\;OO O ?< ?/"N ?/-y/Y vO SO

I r*^\ SO

?I?pq?i?i ? |\? NO hJ

ii

iii?^n-1-1 ?/-\ rvl <N 1^ h>'h^j

_:5:?i=rrhi^?^=rr^rlS~N~i"iiri?Slf^h^.sh

a?Pip?SrFlrjYhrr?

.0-0-:

1 i fs

<
^ 6-1111111(22) 0 Civil
? Buildings

5-n-D p

i?

Churches (57) Synagogues (5)

-p" j

' :I i ::::::=:::::::::::::llrIr::TT:::fl^:::
v o rn

MlI'M I

hi: :

11 111 111 11 111 11 11 i 11 111 11 11 11 11 111 11 11 i j r 111111 ? t 11 it i ri 111 j oio oooooooooooo oooooooooooooooooo 00 aO?-NM"ttft\0N000?O Ot-NWtinONCOOiOi-NMtW^N VO tiAininmihirtininininO <<f lf\tr\tr\tt\tritr\tt\tf\tf\tr\^*$^^^t^^'>$

LLL ILJ ?LJl L I LLl -_l? 11 i 11 iL LIU L^MU-UM^Bbl^MbiLtLlLlli 111 11 111llLu i i ji11 111 111 i i i 11 11 Iii i 11 11
oo ooo o oooo oooo oo ooo m *t m vOfv 00 O* O f- <NM <f irtO N 00 C\ O CN tO NOMDM0 \0 M0 VO? N N N r?s f-sfs N N N N 00

Chronology

therefore more

the notion of a recently advanced civic build sixth-century erosion of pride and standards of living and and are These are inextricably linked with economic change ing. Wolfgang Liebeschuetz has an to traditional scholarly notions of closely yoked in crisis that rendered itvulnerable to Muslim society to Liebeschuetz, and Persian invasion

Fig. 20 Inscriptions in Israel 340 790. From L. Di Segni, "Epigraphic Documentation Provinces 4th-7th on Building in the and Arabia, of Palaestina in The Roman and

c,"

early Byzantine takeover. According

a number of afflictions, such as

Byzantine Near East, vol. 2, Some Recent Archaeological Research, ed. J.H. Humphrey (Portsmouth, R.I., 1999), table 4a.

north Syria, depopulated plague, earthquakes, which began to decline sharply after 550.He additionally argues that the southern cities of Syria-Palestine also declined, though later and less severely. He further appears to hypothesize a slow descent from

the fate of the villages linking of the hill country of northern Syria with the fate of the cities towhich But his view, shared byHugh Kennedy, that the region they belonged. declined from the mid-sixth century, is contradicted by the material Liebeschuetz the importation of and requires reassessment.31 At Antioch wares uncovered in the Princeton mainte red expedition, the Coptic con nance of late into themedieval antique buildings period, and the tinued mention ofAntioch as a town by theArabic geographers argues evidence

Abbasid or Fatimidperiod. after 750, finally hittingbottom in the

is certainly correct in

the Early Byzantine

period with population

and prosperity declining

for continuity of settlement there. In the hinterland of Antioch at Al red ware pottery and coins of Justin I, Justinian, the Coptic Mina, to a and Herakleios occupation phase sixth-seventh-century belong of the site, suggesting continuity at least to theMuslim conquest. As

H. Kennedy, "The Last Century of Byzantine Syria: A Reinterpretation," ByzF io (1985): 141-83. 31

BYZANTINE

EAST

FRONTIER

ECONOMY

237

Foss notes, Chalcis

the last inscriptions from the Limestone Massif hinter land of the city date to 610, and those around Epiphaneia (Hama) and to 605.32 The latter in region, particular, offers a considerable

to briskbuilding activity after$$0 (fig. 20).33 body of epigraphy suggest

In her reassessment of Dehes, Massif Limestone the only published excavation from the concludes that the houses of the villages, Magness to con settlement were built during the mid-sixthseventh-century struction, which calls for reconsideration of the accepted chronology Recent survey in the Amuq of the environment to our valley has added substantially of north Syria. Previous archaeologi

of thehill villages and ofAntioch itself.34

knowledge cal work has generally focused on the limestone hills around Antioch. a data By contrast, the Amuq Valley Regional Project has produced set for the richer as well as the more agricultural lowlands marginal environments. In his of the project, Jesse Casana publication upland noted that Antioch was

far larger than we have previously allowed, because previous work has missed of urban habitation large portions visible in satellite imagery. The plains around the city have also finally

reveals a densely settled, highly interlocking rural network, Amuq with extensive communications and intensive agricultural develop not least ofwhich is a network of canals that utilized thewaters ment, of the Orontes.35

to remains. Until now give up their Roman-Byzantine period begun these sites were overlooked when survey centered on the impres sive tells of the Antiochene hinterland. The landscape survey in the

TheByzantine-Muslim Frontier from theSeventh to theTenth Centuries

The old frontier along the Tigris and Euphrates proved untenable in the face ofMuslim expansion. As the Byzantines retreated from Syria the borderlands shifted farther west, along and upper Mesopotamia, and were, like all medieval the Taurus and Anti-Taurus, frontiers, a broad zone of contact and contention. Muslim Anatolia were both recovered attacks on the cities of serious and sustained. Yet many of these centers Others, somewhat, and few suffered total abandonment.
JMagness, The Archaeology of the Islamic Settlement inPalestine Early 34 (Winona Lake, Dehes Finds at Al Mina 2nd series 37 Indiana, 2003), 196-209; excava a remains the only published 35 J-Casana, "The Archaeological Landscape of Late Roman Antioch," in I. Sandwell and J.Huskinson, Culture and Society inLater Roman Antioch (Oxford, 2004), 102-25.

32

Foss, "Syria inTransition," 261; (Umm el-Jimal) J. Betlyon and B. DeVries, Umm eljimal: A Frontier Town and itsLandscape (Portsmouth, R.I., 1998).

33

A. Lane,

"Medieval

tion from the Limestone Massif, weakness "Dehes

Syria," Archaeologia, (1938): 19-78; L. Woolley, "The Excavations at Al Mina, Suedia 11," JHS 58 (1938): 133 70, illustrating finds of LRi amphorae Byzantine lamps.

inNorth

glaring that archaeologists must move to address (J.-P. Sodini, G. Tate, et al., (Syrie du Nord). Recherches (1976-1978). I?III Campagnes sur l'habitat rural,"

and

Syria 57 (1980): 1-304.

238

MICHAEL

DECKER

remained and Tzamandos, (Sivas), Tyana, Mokissos, centers of size and importance late into themedieval varying period.3 Euchaita, on the northern fringes of the plateau, prospered deep into the seventh century and continued to have a vibrant commercial life like Sebasteia centered on the cult of Saint Theodore After the battle ofMantzikert much of the Anatolian Teron.37 swiftly overran in 1071, the Turks

plateau, whose fate has been largely attrib uted to the nature of the landscape itself and the nomadic element of culture. Since land steppe with the region is naturally viewed of how of the core of Anatolia were large parts high a harsh climate, few trees, and scant surface water, as the domain

Turkish

thus an element of environmental

of pastoralists. There is determinism present in our views

the Seljuks came to control the uplands, while the Greeks were confined to the coasts, a situation not primarily altogether dif ferent from that which prevailed during the Achaemenid empire.38 Neither

nor Cilicia offers the same range of archae Cappadocia as does Syria-Palestine. The Konya Plains Survey, led ological data Baird, explored lands just to the west of Cappadocia by Douglas and found an early Byzantine (sth-7th centuries) peak settlement. this time, numbers of settlements increased substantially: 90 During siteswere still in the percent of the Roman occupied early Byzantine sites contained evidence and 70 percent of Early Byzantine period, of Roman Baird's results thus indicate strong period occupation. persistence

of existing communities. invasion of mar Agricultural is on the land. clearly evidenced, indicating pressures ginal landscapes thus engendered a Demographic growth shortage of high-quality arable land, which led to colonization of alluvial fans and poor hill side soils and their exploitation by crops and flocks. At the same time that the communities of the Konya plain increased in number, they often expanded in size, sometimes dramat

In themedieval population. period (post-7th century) the total area of settlements was only about 100 ha, far below even that of the Iron Age. these data come from only one survey on Obviously the plateau, and many more studies are needed before we can begin
Studies in Honor (Malibu, Cal., 38 M. Hendy Milton of V. Anastos

so. When considered in light of the site numbers, ically burgeoning this increase in inhabited area offers persuasive for a marked proof increase in

36

(Tyana) D. Berges and J.Nolle, Tyana:

Archdologisch-historische Untersuchungen zum sudwestlichen Kappadokien (Bonn, 2000), 517-20, on a 10th-century ecclesiasti cal fragment found at Tyana; (Mokissos) 13 (1995): 109-29; A. Berger, (Mokisos),"

AraSonTop 14 (1996): 27-41; A. Berger, "Survey inViran^ehir (Mokisos)," AraSonTop 15 (1997): 219-37; A. Berger, "Viran?ehir eine byzantinische Stadt in ZsfMzff 48 (1998): 349-429. Kappadokien," (Mokisos), 37 F. Trombley, "The Decline of the Seventh-century Town: The Exception of Euchaita," in S. Vryonis, Jr., ed., Byzantine

1985), 65-90.

A. Berger, "Survey inVirans. ehir (Mokisos)," AraSonTop "Survey inVirans.ehir

(Studies in theByzantine Monetary Economy c. 300-14S0 [Cambridge, I985]? 37-I45) offers an especially detailed analysis of the geographical factors in play. and climatic

BYZANTINE

EAST

FRONTIER

ECONOMY

239

the survey's comparisons with any certainty. Nevertheless, intensive methods and careful attention to off-site landscapes have material of considerable comparative value. From itwe yielded gain to into trends in the highland plains perspective general adjacent we could where expect a similar pattern.39 Cappadocia, The evidence is of a different variety and is sparse Cappadocian by comparison with other parts of the Levant. To begin with, was never urbanized in the same way, nor to the same Cappadocia as much of the eastern empire. For centuries large portions of degree, to the Roman I and II the plateau provinces of Cappadocia belonging I, Caesarea comprised imperial lands. The metropolis of Cappadocia Mazaca was the a hub of communi (Kayseri), region's central place, in part on cations and trade. Caesarea's also depended prominence a as a in the its enhanced standing religious center, position greatly fourth century by Saint Basil. The city continued to produce promi as the tenth nent thinkers into the middle Byzantine period, such century archbishop its communications Its economic Arethas. Caesarea's military

to make

importance lay in and position as an aplekton (imperial field camp). center character iswell attested; the citywas a medieval often congregated. Although in 726, three centuries of in 611 and theMuslims merchants

of commerce where Muslim it fell to the Persians

sack by the Turks in 1092. relatively security passed until Caesarea's From textual sources and the fieldwork of the team of the Tabula Imperii Byzantini (TIB), more than 330 sites are known to the old Diocletianic from that provinces of region, corresponding generally I. In late antiquity, these provinces I and II and Armenia Cappadocia col cities. The twenty-one centers that A. H. M.Jones possessed few lected in his study of the cities of the east included three sites listed as

Several others inHierocles: and Mokissos. Podandus, Doara, regiones and Ciscisus between Kayseri and Tabia) like Camulianae (located were (modern Yaylacik south ofKayseri) apparently small settlements. These latter typify many sites in Cappadocia, which, though they are neither described in detail appear inwritten sources, occasionally to the sent its nor Camulianae bishop yield extensive material data. church councils of Constantinople II and III, and Ciscisus was

rep II. Both Camulianae and Nicaea Council resented at the Quinisext and Ciscisus offer scant material remains for inspection, though the latter's immediate environs contained a number of rock-cut dwell

than to a account, more nineteenth-century ings and, according as domestic structures. At are better read perhaps forty churches that Ciscisus Hans Rott reported the remains of a built cruciform church that measured amodest

39

D. Baird, "The Settlement Expansion

m wide at thewestern arm m 16 long and 4.8 of the transept, yet nevertheless had substantial traces of decorative traces of these installations had been destroyed pretension. Nearly all

on the Konya Plain, Anatolia: 5th~7th W. Bowden, L. Lavan, Centuries, A.D.," in and C. Machado, eds., Recent Research (Leiden, on theLate Antique Countryside 2004), 219-46.

24-0

MICHAEL

DECKER

when

survey team arrived in the area in 1973. The example at Ciscisus many church buildings have survived notwithstanding, more substantial construction and to the present day because of their the TIB

focus of community life.40 are therefore useful Churches proxies for settlement.41 They fre indicators for nearby settlements and thus quently offer chronological add significantly to our view of regional site distribution. Several sev churches are known, including examples enth-century Cappadocian on a number of churches, from Akhisar, where work has progressed to the elev including the famous (Janh Kilise.42 The latter belongs

enth century, as do eleven other churches or monastic installations at Akhisar. era Six others were built during the middle Byzantine There is one sixth-century example, a masonry (8th-nth centuries). church once decorated with murals. an unbroken Other suggest occupa examples Cappadocian tion. At Avcilar, in north of Nev?ehir, the just Rocky Cappadocia at Durmu? rock-cut "Buried Basilica" Kadir Kilisesi dates to the sixth century, the church of Mezarlar Alti Kilise has traces of sev

enth- and ninth-century has images, and the church of Karjibecak been assigned to the ninth century. As at Akhisar, the eleventh

at Avcilar, with seven churches century was a prosperous period a 40 km radius economic activity. Within attesting considerable a more indicative of of Caesarea, and the pla region Cappadocia teau at twelve locales provide material, mostly built masonry large, churches, six churches that attests early and middle Byzantine settlement. The on Mt. around Goreme all Argaeus (Erciyes Dagi), to the fifth/sixth centuries, witness considerable social and vibrancy hinterland of

dated

economic

in late antiquity. Fifth-/sixth-century churches in the immediate of Caesarea are found at Pesek

the metropolis

The (Byzantine Manda). Manda has a polygonal apse, highly decorated with carved exterior pilasters and cornices, and thus bears traces of considerable architec one km to the southeast, were once tural pretension. At Uskiibu, just the remains of a Byzantine to Saint church dedicated and George cave with associated The church of the dwellings chapels. Panagia
40 and M. Restle, Tabula Imperii vol. 2,Kappadokien Byzantini, (Vienna, 1981), 197-98; 206: 328 siteswere recorded, but some are known only from textual evidence and are not localized; (church at Ciscisus) H. Rott, Kleinasiatische Denkmdler aus Pisidien, und Lykien F. Hild 41 J. E. Cooper, "Medieval Cappadocia (9th tomid-nth Century) and the

and Skupi-Uskiibu at cruciform church of the Forty Martyrs

Evidence" Oxford,

Byzantine Elite: The Archaeological (unpublished DPhil thesis, 2002), 57-61. On churches in generally, see

byzantin), 2 vols. (Paris, 1925-1942); N. Thierry, Haut Moyen-Age en Cappadoce: Les eglises de la region de ?avusin, 2 vols. (Paris, 1983-1994). 42 R. Ousterhout, A Byzantine Settlement 2005), inCappadocia 17-78. (Washington, D.C,

Cappadocia

Pamphylien, Kappadokien, (Leipzig, 1908), i73ff.

G. de Jerphanion, LesEglises rupestres de (une nouvelle province de I'art Cappadoce

BYZANTINE

EAST

FRONTIER

ECONOMY

241

at Tomarza,

a cruciform, four-towered church, was built in the sixth

a 30 km north of Caesarea, century. About seventh-century church struc has been found at Hirka, but no eighth-century ecclesiastical ture from this area has been recovered. once Activity is detected again in the ninth/tenth centuries with the rock-cut church near the vil lage of Taforen/Agirnaz thirteenth-century work The and then in the tenth (Byzantine Aragena) in the church of Saints Cosmas and Damian

at Develi.

to the VII Constantine assigned it, along with Mokissos, where Bardas Skleros, a powerful magnate theme of Cappadocia, of the regions most influential family, had a fortified dwelling near in themiddle centuries was unclear. Tyana, but the fate of the city itself until long after The city continued to be an ecclesiastical metropolis unknown. stated in the lapse of Byzantine control. Tyana's status as ametropolis, the ecclesiastical notitiae, finds some support in the recent discovery of church components of a probable tenth-century date. These finds to some extent theMuslim suggest that the city recovered following sack in 831. They century are corroborated and somewhat an bishop, Eustathios, bishop, Leo.44 The other metropolitan by eleventh- or a seal of a tenth twelfth-century

was a sizeable city Tyana in southwestern Cappadocia probably the late Roman period; itwas the capital of the province of during II. Its fortunes during the Byzantine period are largely Cappadocia

in the eleventh cen church at Ispidin was decorated same time the monastic tury and about the complex alleged rock-cut was excavated.43 at Kepez

was seat in late antique Cappadocia lists this settlement, which Mokissos lay (Viran?ehir). Hierocles between the Byzantine city of Koloneia (Aksaray) and Tyana, midway atMokissos to as a there was uphrourion, Prokopios, regio. According or fortified center. Justinian demolished the old fortress and built on level lay ground. also relates that the emperor then built "many churches Prokopios and hospices and public baths and all the other structures that are the mark of a prosperous city."45 The remains of the town presently cover 45-50 ha scattered over four hills flanking the road (fig. 21), vast to from Koloneia major comprise the Tyana. Houses (Aksaray) are modest in size, of the visible remains of the city, and these ity a new one to the west of the settlement, which
43 (Church of Panagia, Manda) Denkmdler, 192-99; and Rott, Constantine VII Porphyrogenitus, De (Vatican, 1952), "Medieval

Kleinasiatische Restle, TIB 44

thematibus, ed. A. Pertusi 65; (seals) Cooper, 378-79 45

(churches ofManda J.Darrouzes,

and Uskiibu) Hild ed., Notitiae episco

Cappadocia,"

2, 228-29.

patuum Ecclesiae

Constantinopolitanae Tyana, 517^;

trans. Prokopios, Buildings 5.4.17-18, H. B. Dewing (Cambridge, Mass, 1971).

(Paris, 1981); Berges, J.Nolle,

242

MICHAEL

DECKER

' ST0MBS
Fptoi road axes
_._ ^^ I

fc^mvir^?^
/ ^^^p-^^
\ ^^s^_/?'

N
'100
\_I I /_/

-'

( l v200
/ 1_^_I

300

Fig. 21 Mokissos. "Varia Cappadocica," (1997): fig. 17.

From E. Schneider Equini, Archeologia Classica 49

BYZANTINE

EAST

FRONTIER

ECONOMY

243

10 20 30 40

50 100

^^^^

^^^^^^^^<X

J\l

m on a side, 10m dimensions are not uncom generally 5-7 though mon. Walls are courses of uncut local commonly constructed of dry sometimes with a rubble core, but rare stone, examples of well-cut blocks used are can be found. Most of the dwell dwellings in and many of the domestic ings irregular shape and windowless, exhibit various types of masonry work and complexes likely multiple in domestic

Fig. 22 Mokissos Equini,

acropolis. From Schneider "Varia Cappadocica," fig. 3.

phases of construction. Since the building techniques of the domes tic architecture are mixed, often within the same unit, even a crude chronology construction is difficult to elaborate. is markedly tend to be built of ashlars and with greater care. at the northwest The acropolis hill (fig. 22), lying 120 m long and 30-50 m wide Further, the technique of house different from that of the churches, which of the site,mea

and is surrounded by a a It is crowned proteichisma (outer fortification). by fort constructed in masonry with three entrances, the main one pseudo-polygonal on the western two semicircu of which lies is flanked edge and by are buildings within the acropolis perhaps barracks belonging to the Justinianic but a group of houses period, encroaches on the southwestern portion of the indicat proteichisma, of or conver continued after the abandonment ing that development sion of the fortress. The date atwhich the latter occurred is unknown, lar towers. A number

sures about

but itmust

are clear certainly be later than the sixth century. There structures atMokissos. sixth- and seventh-century ecclesiastical Of and Bell noted, and Berger further studied, Kemer

these Ramsay

144

MICHAEL

DECKER

mea Kilisesi andKara Kilise. The formerisa smallcruciform church

x 5.5 m at the a suring 5.5 crossing, topped by cupola, while the latter to the is springing of its arches and is constructed of ashlar preserved Restle proposed that itwas originally three-aisled, blocks; though later reconstructions the side aisles. Like demolished apparently

atMokissos, Kemer Kilisesi and Kara Kilise nearly all the churches to late in their case the fifth-sixth centuries. Some antiquity, belong 2 and 3, in the construction of Churches later activity is detectable built, according to Berger, sometime shortly after 600. core of the Justinianic at Domuz do Only outside the city Diizliigu we view clear traces ofmiddle Byzantine building activity in the form to the tenth-thirteenth centuries.46 of Church 23,which belongs Mokissos which were

is one of a hill town that preserved pre-Roman layout, including stepped roads designed to building carry animal traffic to the higher ground. Justinian's architects, who transformed the place, are markedly perhaps responsible for the of paved road found there, but fragments they apparently neither the material remains forms of and

is impressive in its size and in its potential for the study of Anatolian architecture of the Byzantine period, especially domes tic features. In the absence of stratified excavation, the impression of

apparently dispersed further,migrating to the surrounding hills away center. from the by-then decaying Justinianic Middle-Byzantine settlement certainly persisted, evidenced period by the castles routes at Kecikalesi that guard the communication and Comleci (Byzantine Koron).47 The easternmost (Eski Malatya), was Melitene city of "Greater Cappadocia" where limited excavation and surveys associated the scene of fierce encounters

a nor did sewers or a water imposed regular plan they provide supply, both ofwhich the settlement lacks. The middle settlement Byzantine

Melitene

with theGiineydoguAnadolu Projesi (GAP) have recordedremains.

was itself and Muslims.

between Byzantines In 656/57 (d. 680), while still governor of Muawiya the city,which remained one of the border Syria, conquered regions from which theMuslims launched their annual raids. The (thughur) al-Malik, but city fell to the Byzantines during the reign of Abd was rebuilt the Muslims the reign of Hisham b. Abd by during al-Malik. Only in 934 did the retake the city,which the Byzantines forces of Kourkouas devastated. The Byzantines used thoroughly
(Mokissos) E. Schneider et al., Classica and Kara 49 Viransehir 219-37. 47 Hild and Restle, TIB 2,135-37, 216; Schneider et al., "Varia Cappadocica," 141-43. (Mokisos)," ArSon Top 15 (1997):

46 "Varia

Cappadocica."Archeologia (1997): 105-8; (Kemer Kilisesi

Kilise) M. Restle, Studien zurfruhbyzanti nischen Architektur Kappadokiens (Vienna, 1979), 1:46-48; A. Berger, "Survey in

BYZANTINE

EAST

FRONTIER

ECONOMY

245

to resettle the area, which became the Syrians seat of a a strategos and then katepano. Bardas Skleros held the city his revolt against Basil II, but the center remained in briefly during anti-Chalcedonian the imperial orbit until the battle ofMantzikert.48

The site ofMelitene

in the middle Justinian's fortification walls, with latermodifications in the hinterland of Byzantine period, but little else. Fieldwork Melitene ment has provided evidence that elucidates the nature of settle and human exploitation of the landscape. Eugenia Schneider excavated the nearby hill village of Arslantepe, where the Equini are built of double domestic quarters of the Roman-Byzantine period courses of fieldstone with ods similar to some of the houses buildings problem Abbasid coins a rubble core, and display building meth at (Mokissos).49 These Viranjehir show no obvious signs of destruction and, aside from the

(Eski Malatya)

still preserves traces of

of a lack of Umayyad-period pottery, unstratified finds of ware and a trickle of and later material, including sgraffito that generally date to around iooo, suggest that occupation into the eleventh fort upstream conducted in preparation century. Farther afield, for the Karakaya Dam Project ofMelitene at

of the site continued well excavations while recorded traces of a Roman

man) Classical

or less on both sides of evenly the local tributaries, suggesting once the Euphrates, along centers than in peri again dispersed settlement and perhaps smaller ods with fewer numbers of sites. as well as The in remains unsatisfactory data-set for Cilicia archaeological the regards, and survey of the countryside limited. Following

?em$iyetepe, detected post-Classical survey work by Ozdogan (i.e., post-Ro wares at 55 sites, with 22 deemed "Classical." Post compared settlement was found more

in earnest in the began on former Byzantine early eighth century. Their efforts focused and al modern Turkish Misis). Baladhuri Massisa, (Ar. Mopsuestia from the land removed the population Tabari report thatHerakleios the garrisons from and withdrew and Antioch between Mopsuestia in 682, and Arab colonization of Cilicia this area theMardaites these regions. Subsequently frequently raided the between the empires. Nevertheless that became a no-man's-land made stubborn efforts to strengthen and early Abbasids Umayyads which became a staging point for Muslim and colonize Massisa,

many circa 637, both Arab capture of Tarsus and Anazarbos (Anavarza) from the mid contested Cilicia and Muslims the Byzantines hotly seventh century onward. Though Tarsus was apparently reoccupied sometime after the reign of Herakleios, the Muslims destroyed it

48

C. Foss, "Melitene," ODB

2:1336;

E. Honigmann, "Malatya," EP, 6: 230; Hild and Restle, TIB 2, 233-37. 49 E. Schneider Equini, Malatya II. e le Rapporto preliminare delle campagne 1963-1968. Il livello romano-bizantino testimonianze

In 965 John raids against the Byzantines who held the highlands. and the town later Tzimiskes finally captured Massisa/Mopsuestia the city was under the into Crusader control. Subsequently slipped

islamiche (Rome, 1970), ioff.

2.46

MICHAEL

DECKER

sway of the empire of John and Manuel Komnenos until itwas finally lost to theArmenian kingdom of Cilicia. A spectacular
of late antique date is the only noteworthy trace of the early Byzantine architecture.50 awaits remains inhabited today and Anazarbos Since Tarsus of Cilicia the archaeology of the former metropolises excavation, church mosaic I and II is sparse to the extreme. From written sources it is known that Tarsus boasted an important synagogue, churches dedicated to Peter and Paul, and extensive suburbs. But apart from the finds of an early Byzantine cistern, the remains of the late Roman/early

at the southwest of the old city, and Justinian's bridge Byzantine gate over the Cydnus, these Byzantine features have vanished. Stray finds of mosaics, in 1948, and such as the well-known example discovered the reports by earlier travelers of further remnants of the Late Roman an area of at least 70 ha, do littlemore than citywall, which enclosed confirm that Tarsus was excavations a on Golii Kule, large and important late antique city. The in the southeastern portion of themodern

conducted work in the 1930s, offered village when Hetty Goldman to the little attributed and settle reoccupation period of Byzantine a ment of the tenth-twelfth centuries, though probable dyer's work shop may date to around in Tarsus this time. Recent have unveiled a rescue excavations at Republic Square Roman-early street; substantial numbers of late Roman amphora paved LRi, LR5, and LR6 sumption and on the Cilician Byzantine

(LR) types into local and con provide insights production in the city. and Picon noted LRi kilns at Tarsus Empereur coast at Soli and Aigeiai Tarsus will join Antioch ofwine, (Yumurtalik), and with as a and Apamea major oil, and other regional products con Gozlii at Tarsus is

fuller publication consumer producer and tained in these jars.51 The medieval

cursorily cream and Umayyad published, Slip wares across the around the city provide evidence of occupation wares identified as Muslim The green Conquest. glazed Umayyad are more Abbasid these seemingly form the bulk of types; probably but the Phocaean Red wares material recovered from the medieval included plentiful finds of Duochrome eleventh-century date, and also later monochrome
E. Honigmann, "al- Massisa," EI2 50 774a; Hild and Restle, TIB 2, 351-59. 51 H. Goldman, to Cilicia, "Preliminary Expedition 1934, and Excavations at Gozlii 1935,"AJA 39 (1935): 526-49; "Excavation at Gozlii Kule, H. Goldman, Periods

material

from Hoyiik

layers. Later occupation layers ware, suggesting an sgraffito glazed examples.

orientale," inAmphores romaines et histoire economique: dix ans de recherche, ed. and C. Panella (Rome, 1989), 223-48.

Excavations

at Gozlii Kule, and Roman

Tarsus, vol. i, TheHellenistic

(Princeton, 1950); (Republic Square Excavations) C. Toskay, personal communi cation, 15June 2005; (kilns) J.-Y. Empereur and M. Picon, "Les regions de production imperiales enMediterranee

M. Lenoir, D. Manacorda,

Kule, Tarsus, H. Goldman, Tarsus,

1937,"AJA 42 (1938): 30-54;

d'amphores

BYZANTINE

EAST

FRONTIER

ECONOMY

247

preserves an extensive circuit wall of early date, an amphitheater, theater, and extensive Byzantine-Islamic with both built and rock-cut tombs, to name necropolis sarcophagi but a few of the Classical and Roman-period features. The impressive at Anazarbos, which still survive in places, and these undoubtedly continued aqueducts to function at least into the seventh century. To the Early Byzantine three large churches, including the three-aisled Church period belong

as well as Some of the latterwere possibly of Byzantine manufacture the output of Levantine workshops.52 The scant urban settlement record is remedied a little by a glance

in the southeast, dated to 516, is another basilica, also of impressive x 25 x 30.5 located a bathhouse dimensions m). Gough (40 (43 m m), possibly of Roman origin, in the city center, but the presence of or attest to the contin repair. These buildings antique construction ued prosperity and size of the settlement in the sixth century, and we to flourish

a basilica church of the Apostles, 56.2 m x 28.1 m and measuring The second basilica, in the of Roman-period built in spolia. large part southwest of the city,was cruciform in shape (51.8m x 37.3m), while

banded brickworktypicalof early Byzantinebuilding indicatesa late

presume that the town continued Muslim invasion.53 must

to the right up

the site, including the cur fortification works around Byzantine tain wall atop the spur of the acropolis, attest to the return of Greek rule. On the northern flank of the fortified spur, Gough discovered Justinianic-period reworked. Given cal examination, Byzantine, Muslim, fortification

A modest middle Byzantine shrine, decorated with frescoes still lies on the shoulder of the acropolis visible during Gough's visit, hill, on the path leading to the fortress itself.Traces of the middle

was later activity that apparently the fact that the city, despite limited archaeologi remains from the early has yielded considerable middle

and Armenian occupation Byzantine, remained an important there is no doubt that Anazarbos periods, the early medieval period, supported as itwas by place throughout its and extensive territory that, in the tenth century, superb position remained optimal for both agricultural and herding.54 century, the Byzantines evacuated the and Antioch, between Tarsus, Alexandretta, they appar population sections of the ofMopsuestia elsewhere, people ently transported large When, in the mid-seventh
52 F. Hild and H. Hellenkemper, 5:Kilikien Tabula 54 M. Gough, "Anazarbus," AnatSt 2

Imperii Byzantini

und Isaurien

(medieval ceramics) (Vienna, 1990), 428-39; F. Day, "Islamic Finds at Tarsus," Asia 53 (March, 1941): 143-48. F. Hild and H. Hellenkemper, TIB 5,

(1952): 85-150; al-Ist.akhri, Kitab al-Masdlik 2nd ed., ed. M.J. De Goeje wa-l-Mamdlik, (Leiden, 1927), 63; Le Strange, Palestine under the Muslims, 389 (above, n. 22); phases) (lower citywalls, medieval

H. Hellenkemper, von Anazarbus,"

"Die Stadtmauern inXXIV Deutscher

Orientalistentag vom 26. bis 30. September 1988. Ausgewahlte Vortrage, ed.W. Diem and A. Falaturi (Stuttgart, 1990), 71-76.

178-84.

248

MICHAEL

DECKER

perhaps Pedias.

to areas of the

in thusbecameArabicMassisa, the majorMuslim stronghold Cilicia

plateau.

The

former Byzantine

Mopsuestia

retains Like the cities of Tarsus and Anazarbos, Mopsuestia scattered reminders of its late antique and medieval past. Some only were maintained remnants of the Roman in city wall survive that the Early Byzantine era; they enclosed an area of about 40 ha. There is some about which rulers were responsible for the disagreement reconstruction of the city II is known rule: Umar during Muslim a cistern, provided the suburban complex of Kafarbayya with traces of which are and Hisham apparently added the fortifications, survey in the early 1950s covered most of low land Cilicia, but primarily explored tells and sites with archi larger tectural remains and thus offers limited scope for quantification. to have

visible today. Seton-Williams's

a restricted corpus of was available for Only comparative materials the analysis of finds from this work and the resultant chronologies are therefore rather a occu coarse-grained. These data suggest peak in pation in the Roman-early Byzantine periods, but offer little insight ware into the crucial Islamic transition. Finds of Byzantine-early sgraffito at sites argue for a wide distribution of settlement thirty-nine

from the eleventh century onward.55

SettlementTrends: The Survey Data

the limitations of the survey data are Despite presented above, which functions of in part the in part the condi methodology employed and tions inwhich they were undertaken, one can venture general obser vations on broad settlement patterns from the sixth to the eleventh centuries. Along the Tigris and Euphrates corridor, all present indica

portion of the Tigris watershed considered here, the Su suggest intensive habitation valleys of the Batman Su and Garzan late antiquity, which fanned out across the over a during landscape, variety of soil, indicating substantial pressures from the population on available land resources. Farther east, along the Bohtan Su River and in the of Cizre, late settlement was present, but plain antique more restricted than itwas farther west and less common than late medieval/early Ottoman period sites. the Along Euphrates, multiple surveys record a very high density of a sites followed in the Byzantine-period by significant decline Islamic period. For the Sajur early example, Survey, which examined sections of the around Manbij large (ancient Hierapolis, landscape

tions point to a high density and wide distribution of settlements in the late Roman/early Byzantine period of the fourth-sixth centuries. From the Tigris, the results remain qualitative and we can gain little more than an impression from them. the tributaries of Along thewesternmost

55

M. V. Seton-Williams,

"Cilician

Survey," AnatSt

4. (1954): 121-74;

(sgraffito chronology and forms) K. Dark, Byzantine Pottery (Stroud, Gloucestershire, 2001), 65-77.

BYZANTINE

EAST

FRONTIER

ECONOMY

249

of Euphratensis), shows Byzantine settlement numbers metropolis even in the Bronze higher than Age. These results broadly parallel those of other surveys in the northern Euphrates valley. Farther upstream, the hinterland of the Roman/Byzantine city of Zeugma witnessed explosive

in settlement. The Birecik-Carchemish survey growth evidence for a decline of sites around the old central-place provides good settlement of Carchemish Colonia Aurelia (Greek Europus/Roman while a

Dura),

at increase corresponding along the Euphrates a translocation of in the Zeugma probably indicates population early states that the Muslim conquests, Algaze Byzantine period. Following the area was population suggests that, since "sharply depopulated."56 Wilkinson rose in the Balikh and portions of the Khabur apparently

picture valley during the early Islamic period, the archaeological to the latter areas.57 may reflect migration Perhaps the Umayyad authorities removed the population from trievulnerable frontier zone, shore. It is they did with exposed coastal cities along the Levantine from possible that the Byzantines themselves removed the population these regions, either as a matter of defensive policy, as Herakleios or in the course of border and the Antiochene, ordered for Cilicia as

The policy of forced population transfer came raiding and capture. to V, who deported large particular prominence under Constantine numbers of people from the Euphrates region and settled them in Thrace. Constantine seized a number of the inhabitants of the district of Kalikala who were were eastern frontier, but they eventually recaptured by theMuslims and settled in Syria.58 of settlement in the early with a Concordant general expansion are the results of surveys from the Euphrates valley Byzantine-period across the river from Samosata, where region northwest of Edessa, subsequently settled elsewhere in the

and Titrij (figs. 18 and around Kurban Hoyiik regional prospection a 19) has furnished evidence for dramatic settlement peak in the early and a general lack of early Islamic settlement until Byzantine period when a slight recovery emerges in the record. the Abbasid period, in the in the plain of Adiyaman bank of the Euphrates, On the right

initial survey results signal a spike in former territory of Samosata, to The lack of settlement from the Roman early Byzantine periods. refinement in the ceramic analysis as presented in the preliminary means that interpretation of reports of the work around Adiyaman the medieval a to must be circumscribed stating that largely period series of ceramics is known, indicating some level of habita

56 Algaze et al., "Preliminary Report," 207 (above, n. 4). 57 Wilkinson, P. Charanis, "Regional Approaches," "The Transfer of Studies in Society 246 (above, n. 2). 58 as a Policy in the Byzantine 3 (1961): 140-54.

unbroken tion whose

cannot be deter to the early Byzantine position relative on the Dam as yet. Similarly, mined Survey Karakaya Ozdogan's we Turkish upper Euphrates can only be crudely quantified; from it was see that the settlement noted under the rubric "post-Classical"

Population and History

Empire," Comparative

25O

MICHAEL

DECKER

The Jazira evidence contrasts with that of a of the large part In several instances from the former region, fewer Euphrates valley. late antique Byzantine settlements were recorded (and Sasanian) compared with the number of those during the later Islamic period. In one

uted to the chronologicalboundaries imposedon thedata.

far greater than other

periods,

a fact that can, in small part, be attrib

than those of the following Islamic period. Several factors be responsible for these variations, but according to the present may data, Byzantine occupation of the northern Jazira was less dispersed not part and less intensive than in other areas of the East. Although of the quantified corpus considered here, Bartl's work on the Balikh River argued periods. One for a settlement and Abbasid peak during theUmayyad is that settlement of the possibility early Byzantine

numerous

instance (Tell Brak) results are inconclusive because sites of CE were in two instances the firstmillennium grouped together, and sites were more (Tell Beydar and the Upper Khabur), Byzantine

was restricted at least in in this part due to the prox region imity of the frontier and the unsettled relations and open warfare that frequently prevailed between Byzantium and Sasanian Persia. are known from both the Although Byzantine-era irrigation canals Khabur and Balikh valleys, it is likely that landowners and the state period were loath to invest in such projects over much of the later sixth and early seventh centuries, when raiding and invasion would have ren dered such investments vulnerable. recovery of this region's population Islamic-period seen is perhaps explained when against the backdrop of settlement after retreat from other areas. Both the Khabur and the Balikh flow cities of and Resafa, where the Umayyad Raqqa a caliphs took special interest in developing canals and other agrarian features. It is probable that the northern Jazira witnessed the arrival of state-ordered transfers or refugees from the perilous Euphrates Further study is needed to Tigris valleys. verify this picture, particu close to themedieval The probable

larly since traditional survey tends tomiss large numbers of sites. Of the surveys in the Jazira, only the Tell Beydar Survey considered non mounded sites, low sites, and other off-site features. Unsurprisingly, the Tell Beydar region, which lies in the frontier Byzantine-Sasanian zone itself (thus the late ceramics are classified "Sasanian"), antique shows a much incident of late antique settlement, since ithas higher
59 (Birecik-Carchemish) Algaze, 1991, "Final Report," 23; (Adiyaman) S. R. Blaylock, D. H. French, and G. D. Summers, "The Adiyaman Report," 124-30; Survey: An Interim (Karakaya) M. Ozdogan, The Lower Euphrates Basin 1977 Survey (Ankara, 1977); (Titris) G. Algaze, A. Misir, and T.J. Wilkinson, "?anliurfa Museum/ University of California Excavations and Surveys at Titris Hoyiik, 1991:A Preliminary Report," Anatolica 18 (1992): 33-60.

BYZANTINE

EAST

FRONTIER

ECONOMY

25I

that Byzantine and Sasanian settle ment in the Jazira tended away from tells and instead the occupied areas of the as I have noted above, tells have been low plains.60 Since, demonstrated elsewhere survey work, a significant number of late our net and the as it stands sites have picture antique slipped through is open to considerable future modification. In theAntiochene and Cilicia, the fourth-seventh century phase the focus of most older left a strong marker in the archaeological tions of the countryside around Antioch record. Previous have, due to investiga

been

methodologi cal weakness, the early Byzantine occu drastically underrepresented of the area. The dispersal of settlement and migration away pation from the traditional, defensible hill sites and into the open lowlands denotes a sense of security and a prevailing need for land. Investment in of routes of communication, irrigation canals, the concatenation into the highlands around the city and an increase in agri all indicate a growing population of Antioch Sometime between the seventh and cultural exploitation and trade. and the extension of settlement the small, widely dispersed settlements that A real decline in num typify the early Byzantine period disappeared. tenth centuries, however,

bers of people is likely:the suburbsofAntioch, highlydeveloped in

late antiquity, were in all likelihood abandoned by the end of the sev l enth century. The precise pace and nature of change in settlement conquest of this part of north Syria remain following the Muslim As noted above, lowland Cilicia in the certainly shared general expansion of late antiquity. The Byzantine more remains mysterious. The limited

uncertain.

and economic population transition once Muslim

at Tarsus, and continued investment in the Umayyad material finds life from and Anazarbos cities of Tarsus, Massisa, signal vital urban at least the Abbasid limited survey strongly suggests that period, and a number of rural sites continued on through themedieval significant did transferwrought by Herakleios period. The proposed population the no the plain. However, not, apparently, thoroughly depopulate man's-land condition of the countryside and frequent sparring of the sector of the eastern frontier no doubt two powers along the southern threatened communities, particularly those proximate to the major

routes of communication. The present evidence argues for strikingly different fates for the zones of the northern frontier between Byzantium and most exposed Islam, and the southern Levant. At the regional level, the landscape
60 Wilkinson, "Regional Approaches," "The Archaeological

237, 246; Casana,

suffered and Samosata of the city territories of Antioch, Zeugma, attrition: the number of sites dropped precipitously from the deep Islamic periods. In other northern regions, such as Byzantine/early remained Balikh and portions of the Khabur, populations along the

104-5 (above, n. 35). Landscape," 61 J.Casana, "From Alalakh toAntioch: Settlement, Land Use, and Environmental Change Turkey" in theAmuq Valley of Southern (unpublished PhD dissertation, of Chicago, 2003), 309-10.

University

252

MICHAEL

DECKER

stable or even increased. Life also continued with the southern Levant at places like Bet Shean

little

disruption el-Jimal.

in

and Umm

action. The extension of Muslim attacks into Byzantine military to flee to were meant that Anatolia pressured exposed populations safer locales. The remains of sites likeMokissos support the argument did not escape the travails visited upon Byzantium the seventh and eighth centuries. But the persistence of church during domestic remains demonstrate buildings and considerable medieval that Cappadocia

no small part Cilicia and Antioch suffered depopulation, in a function of imperialpolicy, but deepened by annual jihad and

in that even poorly defended towns need not have been abandoned the face of theMuslim threat. The substantial numbers of small rural

was either in sites argue that any decline Cappadocia population or In fact the rural face of settlement and evidence slight ephemeral. over sizeable areas of the for plateau defy the agricultural exploitation a lack of cities connotes a low traditional view that population.

theFrontiers: Settlement and Agrarian Farming Eastern Borders Life along Byzantium's
The world

arena were both a economic the late antique Mediterranean high a in the movement volume of products and of those velocity high products. Bulk, low-valued goods traveled both overland and by ship by degrees and distances rarely witnessed Greek cultural influence in the Levant. the Byzantine Empire, heart of Asia Minor, we see material

of Early Byzantine agricultural production and trade was highly integrated, with exchange of imported goods penetrating deep even in remote landlocked areas. Hallmarks into the of countryside,

in the thousand years of In the frontier regions of no less than in the more secure cities at the popu simply of the by an

full occupation of the countryside is a feature of the early era eastern Byzantine provinces, especially throughout many of the corridor.62 large stretches of the Tigris-Euphrates The most obvious example of the integrated agrarian economy era comes from itswest of the frontier zone of the early Byzantine equally ernmost extensions, where Antioch or were and its vast hinterland, whether with mountain settlements. Here, where plain, packed archaeological work has been varied and sustained, the multifarious are an pieces of the puzzle impressive picture of heavy rural building settlement in symbiosis with one of the true of the early metropolises

a signs of generally dense lation and vigorous agrarian life. This agrarian realm was not self-contained, but in all cases tied to the major urban centers rural districts. The impressive attended degree of urbanization

data attest the further advance of Byzantine world. Palynological agri culture into marginal regions and particularly the dominant place of

62 M. Decker, Tilling the Hateful Earth: Agriculture and Trade in theLate Antique East (Oxford, forthcoming).

BYZANTINE

EAST

FRONTIER

ECONOMY

253

the olive and wine

in the economy.

In addition

to the finds of numerous at

oil

presses, suggestive of major investment aimed a consistent surplus production, potteries that produced iwares (LRi) line the northeast Mediterranean Amphora These kilns produced Syria olive oil, from thousands of farms in the northeastern are found on hundreds sent an economic Farther Euphrates Neolithic inland, of sites around current of considerable the relatively breadth.63 to Rhodes.

capturing late Roman coast from and that

once filled with wine jars, Levant, theMediterranean

and repre

rich valleys of the and Tigris at least the had been heavily exploited for agriculture since era, and it is best to see the Byzantine period within the

and land framework of long cycles of settlement and abandonment use and disuse. Through this lens of the longue duree, the Byzantine moment in Syria and upper Mesopotamia appears as one of substan areas. in many tial growth and dynamism pressures on Population limited arable land encouraged farmers to seek more efficient agricul

other organic waste were carried from villages, into the soil, creating the sherd scatters that first recognized as relics of the deliberate distribution of Wilkinson such organic refuse. Such strategies represent human efforts to sus and vessels containing broken, and ploughed

tural strategies, and towork the land with greater regularity. Shorter meant that more nutrients were extracted from fallowing periods chamber pots the soil by crops with less time for recovery. Ceramic

tain or increase yields by replacing carbon and nutrients normally lost to the soil Farmers thus aimed to stabilize yields, through cropping. poor into which cultivation had been forced by ground intensive strate These pressure, or reclaim wasteland.

ameliorate

population those areas and mirror conspicuously gies attend high population sites are witnessed over a where greater numbers of early Byzantine broad swathe of the middle as well Zeugma-Apamea in the hinterland Euphrates, especially as near Edessa and Samosata.

of

maintenance and investment in the irrigation Early Byzantine at present: the cities of can works along the frontier only be surmised
63 (Presses) O. Callot, Huileries Syria: A Palynological Study," Quaternary International 73-74 (1999): 127-36; (LRi amphorae) M. Decker, "Food Northwest for an Empire: Wine inNorth and Oil Production Syria," inEconomy and Exchange in theEast Mediterranean during Late Reconnaissance Euphrates Archaeological Final Report," 22; (sherd scatters? Project; Middle Town and Euphrates) Wilkinson, Country in South-eastern Anatolia, vol. 1, Settlement and Land Use atKurban Hbyuk Basin and Other Sites in theLower Karababa

antiques de Syrie du Nord (Paris, 1984); (palynological data) U. Baruch and S. Bottema, "ANew Pollen Diagram from Lake Hula: Vegetational, Climatic, and Implications," inAncient Anthropogenic Lakes: Their Cultural and Biological (Ghent, 1999), 631-36; Y. Yasuda, H. Kitagawa, and T. Nakagawa, "The Earliest Record ofMajor Anthropo logical Deforestation in the Ghab Valley, Diversity, and A. C. Roosevelt ed. H. Kawanabe, G. W. Coulter,

Antiquity ed. Kingsley and Decker, 69-86. "The Definition of T. J. Wilkinson, 64 Ancient Manured Extensive Zones byMeans of Techniques," JFA 9 (1982): 323-33; (sherd scatters? et al., "The Tigris Zeugma) Algaze Sherd-Sampling

et al., (Chicago, 1990), 69-79; G. Algaze "The Chicago Euphrates Archaeological Project 1980-1984: An Interim Report," Anatolica 13 (1986): 44.

254

MICHAEL

DECKER

Callinicum, and Circesium Barbalissos,Dibsi Faraj, Sura (Souriya), was or beyond the "zone of uncertainty" inwhich rainfall lay in
unstable. on a number of security thus depended interlocking old canals, farmers maintaining with the Byzantine-era strategies, as that discovered on the Balikh, and excavating new ones. such In other areas of the frontier, as at Hierapolis and farther east at in the Chalcidike, farmers in late antiquity developed al-Andarin Food extensive

that tapped galleries (qanats) drainage underground water. In most instances, though, aquifers for irrigation and drinking cultivators and early Byzantine period dry farming prevailed, with extensive cereal cultivation (generally barley) supplemented in the steppe or herding uplands.65 can be difficult to detect, has Pastoral activity, which rarely of archaeologists. Such an omission from a serious gap in our that has knowledge landscape studies represents been confronted. Generally nomadic elements within only recently occupied the attention

to the the Byzantine-period population would have been confined as the presence of their massive flocks margins formuch of the year, would have wrought irreparable harm to standing crops. As the early increased, the inhabitants sought new farm Byzantine population lands. Thus cultivation crept farther into less desirable, peripheral The limestone hills over landscapes generally reserved for pastoralists. traces of livestock pens looking the Euphrates and theAmuq preserve centuries. In the vil and encampments dating to the fourth-seventh the restricted landscape and lages and hill country around Antioch, activities of the nomadic tightly packed villages naturally limited the

In contrast the Jazira, with its open elements among the populations. range and less tightly bound settlement, made an ideal home for herd and Lakhmids quarreling over grazing ers;we hear of the Ghassanids in upper Mesopotamia during the drought around 539-66 The with hot extensive steppe the Anatolian summers. Most lands of the Jazira share critical features low precipitation, open country, and plateau:

of former Galatia, Phrygia, and Cappadocia receive less than 300 mm of precipitation, while the tree annually less landscape and challenging environmental conditions rendered
fermes byzantines aux palais omayyades, ou l'ingenieuse mise en valeur des plaines inAux steppiques de Chalcidique," origines de I'archeologieaerienne (A. Poidebard and 187s-19ss), ed. L. Nordiguian J.-F. Salles (Beirut, 2000), 109-22; B. Geyer and Y. Calvet, "Les steppes arides de la Syrie du Nord au Bronze ancien ou Ta premiere conquete de l'Est'," in Conquete de la steppe et appropriation des terres sur lesmarges arides du Croissant fertile, ed. B. Geyer (Lyon, 2001), 55-68 ; see also Redford, "Archaeology" (above, n. 18). 66 (Animal pens) Algaze et al., "The Chicago Euphrates Expedition," 44; Casana, "From Alalakh toAntioch," 419 and appen dix D; (grazing dispute) Prokopios, Persian War 2.1, trans. H. B. Dewing Mass., 1914). (Cambridge,

65 (Zone of uncertainty) T. Wilkinson, "Settlement and Land Use in the Zone inUpper Mesopotamia," and Agriculture in Northern ed. R. Jas (Leiden, 2000), Mesopotamia, inRainfall 3-35; (trade) C. Morrisson and J.-P. Sodini, "The Sixth-Century Economy," in The Economic History ofByzantium, vol. 1,ed. A. Laiou (Washington D.C, 2002), 206-12; (qanats) R. Jaubert and F. Debaine, "Des of Uncertainty

BYZANTINE

EAST

FRONTIER

ECONOMY

255

? \ >

'

v" s

> ?'o. or

^\

,-a

J-^--O?'"S.

o >?--?rr

?v,n?

on

lor0'"'

. ,

y^

^%1P|?
V5 5-0/

^^?sV4

V(X

?^J^fe.
\ tr

("-1

xn'-Nc_

7)

~ '

? t

Livestock

ro b e

9)V0y -?o\\ ^ ^?o

?*, ?o

L'v?stock^^__?<__-^^ ^ <T c^r^S"^->A2^ ^ ^*S

N x

l/T^?'^x \ I I ^^-^ '/ ?\N

V \ /V^^jT// **-- N-^y /' 1 / \,'

-^

Schematised 500m. . Contour Line

\r$p> T^_^;
considerable toralism winter to view more amenable to pas plateau than crop husbandry. The extremes of summer heat and cold and a dearth of surface water have influenced scholars areas of the Anatolian
Studies

i.

Fig. 23 Land use, Anatolia. From Hendy, in theByzantine Economy, map 13.

as a hot and cold desert, the uplands of Turkey speck led with cities in a few more favored areas. Over the highlands at least in Cappadocia, ranged flocks of cattle great and small and, horses of exceptional The textual sources make frequent quality. mention the Cappadocian Fortified both by modern countryside. texts like the stance and early-ninth-century Life of Philaretos by as synon the Merciful, many have naturally viewed pastoralism of sheep, goats, cattle, horses, mules, and even camels in circum

a view map ymous with Anatolia, by Hendy's encapsulated (fig. to livestock is in which central Asia Minor 23), completely given Yet the present evidence challenges the assumption production.67 was essen that the nature of herding on the Anatolian plateau Instead I would suggest that the early and middle tially nomadic. Byzantine Jazira, a

was similar to that of the stock-raising of the plateau characteris environmental region that shares important tics with Anatolia. highland It is necessary tomake a key distinction in the kinds of stocking in Anatolia. Rather than nomadism, which implied large practiced or seasonal movements of herds with a mobile human popu annual at the heart of the plateau lay agropastoralism a blend of field involves economy. Agropastoralism cropping and lation in attendance,

67

R- Teja, Organizacion economicay en el siglo IV, segun los social de Capadocia capadocios (Salamanca, 1974), 29-34.

padres

256

MICHAEL

DECKER

set of crops interlocking The and management of Philaretos once more pro techniques. Life vides a useful departure point. The story upholds the ancient para of the rich person whose wealth is tied to the land; the saint has digm since animals only thousands of animals

animal husbandry and differssubstantiallyfrompure nomadism,

are one constituent of an

to an estate may be sent away for substantial shepherds belonging to graze their flocks in or scrub periods of time outlying hill country but the enterprise remains dependent on an estate center, and land, the herdsmen remain under

out over granges, but many of these fanning are must therefore have irrigated and comprised cropland and pos was therefore sibly also watered pastures. Philaretos's stock managed in combination with such conditions, pursuits. Under agricultural

the control of villagers or estates. It is this model that I propose best depicts the land prevailing Byzantine use of the In the surface remains of medieval plateau. Cappadocia in rock examples was from stocking activity managed for example, the rock-cut settlement and underground are known

period mangers and stock pens cut settlements attest thatmuch

village settlements. At Filiktepe, contained several animal pens, and similar installations

at Ovaoren. from Rocky Cappadocia Stables for caprids and equids are known from the sites of Ke?lik B and Soganh Dere i. in Anatolia Mixed and in Syria-Mesopotamia, where farming in limited areas and under less domestic crops were often produced on animals to assist crop pro frequently relied or and to bulk up the diet and resources of the household on the estate. animals pro species and circumstances, Depending vided traction, wool, milk, meat, leather, industrial fats, bones, and than ideal conditions, duction manure Intensive farming held sway over the rural life of the eastern practices regions of Rome and Byzantium. The kernel of these high-input agrarian structures is found inwhat today we refer to as mixed farming: the combination of various crops, or crops and animals. The in medieval stabling of animals Cappadocia signals on the medieval frontier between Byzantium farming practice that and
68 (Filiktepe/Ovaoren) V. Castellani, "Human Underground Settlements in A Cappadocia: Topological Investigation of the Redoubt System of Gostesin (NE 20)," inG. Bertucci, R. Bixio, and M. Traverso, eds., Le Cittd sotterranee della Cappadocia (Genoa, Dere 1995), 41-52; (Keslik B/Soganh Cappadocia, 112-15 1)Cooper, Medieval (above, n. 41).

that served as both

fuel and fertilizer.

Islam fits squarely with what we know to be the dominant forms of as a whole. Confined livestock provided exploitation for Byzantium an source of fertilizer; husbandmen used grain stubble as important litter,which subsequently absorbed the nitrogen-rich animal waste and rendered an important fertilizer source. The maintenance of

soils, particularly the weak and often desiccated earth of the plateau, see these efforts in demanded assiduous care. We practice from the nearby Konya plain, where widespread, low-density pottery sherd scatters of the like that early Byzantine period represent manuring in witnessed evidence for a Upper Mesopotamia.69 Comparative

69

(Stabled cattle) H. Beckh, ed.,

Geoponica

(Leipzig, 1994), 2. 21-22; (Konya Plain) Baird, "Settlement Expansion"; (sherd scatters) Algaze et al., "The Chicago 44.

Euphrates Expedition,"

BYZANTINE

EAST

FRONTIER

ECONOMY

257

mixed Crusader

hill from Gritille, a fortified near Samosata inhabitedonly fromthe

The contested landscape around period onward (see p. 234). old Samosata (Samsat) was slow to recover from the damage suffered centuries of and Muslims. during the feuding between Byzantines

farming

on regime that centered agropastoralism

is known

When

the caliphate invested in the borderlands, they directed their and the creation and bolstering of the efforts toward Cilicia, thugur, to the detriment of their possessions around the Euphrates. One use of Byzantine farmers is vestige of the accelerated land in the rock faces in the remains of thousands of dovecotes Although western world in the present-day the art

present of Cappadocia.

was Eurasia. Indeed pigeons widespread throughout Ages the practice remain an important component of Middle Eastern farming.We may rock smile at the thought of a small creature like the domesticated more than a curiosity, or a nuisance dove being anything perhaps, but the realities of a world without an abundance

Middle of raising vanished, in antiquityand the pigeons has largely

of food were quite dif ferent. The pigeon represents one of the fastest-reproducing forms of on their to humankind. Without constraints placed protein available the initial ten breeding pairs recommended by the reproduction, over 150,000 author of the Geoponika (14.6.9) could have produced in five years. Birds were used formeat; in the early Roman offspring Varro offered advice on aviculture and noted that, due to the period urban market of Rome, bird-raising was extremely profit flourishing able. In late antiquity the price of a single pigeon could command as much as a modius of grain. nearly It is uncertain whether a Anatolia, from the birds was

medieval

in or not birds were widely marketed but inmany ways the meat that could be gained

farm secondary consideration for Cappadocian waste was vital to life. A peren ers. Rather the agricultural pigeon nial problem for farmers of the pre-industrial era, and one frequently cited by scholars as inherent to the primitivism of agriculture in the Mediterranean world, was a lack of fertilizer. Under normal condi tions, especially in naturally poor lands like those in many parts of soil fertility would central and eastern Anatolia, quickly decline. a low return that would stabilize at prob Eventually yields would let alone generate any not subsistence requirements, satisfy ably that the remedy for kind of surplus. Byzantine farmers understood was the the problem of poor soil recovery and low yield application one of the most potent fertilizers of of pigeon waste, undoubtedly the pre-industrial age. To provide but one brief example: a relatively modest flock of 500 pigeons would, over the course of a year, provide nutrients for up to 5 ha of barley land. Three to four ha of enough have been sufficient to sustain the average family. grain land should

258

MICHAEL

DECKER

KFig. 24 Dovecote,

Erdemli, Cappadocia

Pigeons would seeds ofweeds

no care on the margins of the rural settlements and required almost part of the cultivators of the land. Dovecotes

have met most of their feed requirements by eating the and other wild plants. They thus exploited the untilled

were in where many widespread Cappadocia (fig. 24) were rock-cut and thus survive partly today. Demenge investigated many of the pigeon houses in this region; most he found were rather modest, housing thirty or one hundred nesting boxes, perhaps each indicative of a esti breeding pair of birds. Around Goreme, Demenge mated that the fifty-seven Byzantine pigeon houses could have held 120,000-160,000 been estimated birds. More than five thousand dovecotes have to exist area of just in the region of Urgup, an only area alone the tenth century this have had 500 km2.70 During might consideration of any other compo upward of 325,000 birds.Without

nent of the means agrarian economy, these birds alone provided the to sustain five thousand sufficient fertilizer for people by providing 3,250 ha of arable land. neglected and enigmatic symbol of but of a specific type. Agriculture associated with farming, these structures would have been intensive. I have, for ease Although as a fertilizer of analysis, demonstrated the potential of pigeon dung arable Dovecotes are a somewhat

in Cyrenaica covered at Apollonia that lies adjacent to the oil press. the olives that filled the press at one time were Undoubtedly largely the product of the fertilizer that came from the columbarium. These

for a subsistence crop like barley, dovecote waste in all likelihood was and intended for valuable cash crops that were carefully managed we found in and orchards. This practice was gardens typical of what as one dis know of dovecotes from the early Byzantine period, such

70

G. Demenge, "Pigeonniers et ruchers byzantins de Cappadoce," Archeologia 311 (i995): 45

BYZANTINE

EAST

FRONTIER

ECONOMY

259

Fig. 25 Meskendir Valley runoff irrigation system. From Bixio et al., Cittd sotterranee, 281, fig. 2.

structures are therefore well ern

suited to the broken

landscape

of south

where the eroded landscape is fragmented and soil Cappadocia, occurs in such conditions, pigeon pockets of degraded rock. Under waste the landscape and provided an ideal ameliorated considerably solution to the difficulties of sustaining gardens and orchards. are useful proxies in the archaeological Pigeon houses for settlement patterns, in that those villages with which columbaria dovecotes were once associated have vanished, while record

many the rock-cut

more remain. They are, important, invaluable for perhaps an entire that remains largely enigmatic. landscape understanding As indicators of intensive agriculture, these dovecotes are impressive in their numbers. But there are further indicators of substantial frontier zone. Irrigation agriculture elsewhere along the are known, in theMeskendir works valley (fig. 25). There, especially an of large conduits tapped into springs, the flow of impressive array which was augmented by flash floods that rushed down the seasonal watercourses the beds of that formed amid the broken hills. Within investment in these streams, dams and terraces were constructed, but an effective one in the semi-arid landscapes theMiddle East. The systems inAnatolia, while a laborious process, ofNorth Africa and are

not as numerous,

260

MICHAEL

DECKER

we

water so that it into the retarding the velocity of the percolated soil. Several such systems have been discovered and recorded by the Societa Speleologica Italiana. Far more survey work is needed before approach and

of the late antique runoff farming installations spread These farms obtained their water from the the throughout Negev.71 The dams runoff of sudden rainstorms that collected in the gullies. and weirs both diverted the flow of these violent episodes into cisterns,

reminiscent

of the range, precise any understanding chronology, in of these water-harvesting systems elsewhere frequency Given the climate,

Anatolia.

terrain, and the pervasiveness of such systems on three continents during the ancient and water-harvesting medieval periods, there is no doubt that central and eastern Turkey will produce many more such features.72 We have a hint of some of these finds from the letter of Harun to Constantine VI, which described truce: al-Rashid the state of the border

lands at the cessation of the Arab-Byzantine

You knowwell that,throughthe God accorded fidya [treaty],

to each of your categories and each of your classes considerable in several areas. benefit and great these I advantages Among

mention:...The

fact that your laborers and artisans were quick to rework their land and whatever they disposed of...; repair out in order to rebuild and innovate in they spread agricultural the summits of the mountains and methods; they abandoned themarshes and went, in the midst of their

dwellings... digging canals, planting trees, and causing springs to burst forth, in such a way that they prospered. Their situation flourished, and their mountains prevented the wheat from fertile,...whereas today, they are their lands...they have abandoned cultivating fields, the fertile lands, and the water canals for the became

arid mountains.73 Archaeobotanical add another intriguing to the of the agrarian face of Anatolian settlement. At piece puzzle farmers depended heavily on rye, a find that Beycesultan medieval heralds a departure from the norm of tastes, especially Byzantine for the Mediterranean on which region, traditionally relied barley and wheat. it was Rye found few admirers in the ancient Mediterranean; the food of northern barbarians, to and its adaptability environments From Europe. a plays out throughout theMiddle palatability standpoint, Ages rye is a poor finds from farther west

colder, moister in northern make

P. Mayerson, The Ancient Agricultural 71 Regime ofNessana and the Central Negeb (London, 1961);M. Evenari, L. Shanan, and N. Tadmor, Desert 72 TheNegev: The Challenge 1982). ofa (Cambridge, Mass., Bixio, Castellani, 279!?. Byzantium Viewed by 2004), 92-93. (Cambridge, Mass., N. El-Cheikh,

it a strong choice for the not the urbane. While practical, if rye must have been present in Anatolia from early times, its expansion

bedfellowofwheat or even barley, but itshardiness and high yields

and Succhiarelli,

Cappadocia, 73 theArabs

BYZANTINE

EAST

FRONTIER

ECONOMY

z6l

in cultivation departure

era is during the Byzantine again striking both for the tastes that it heralds as well as the from Graeco-Roman cultivators. With the breakdown of

adaptability of the Anatolian urban many Graeco-Roman

at Gritille Survey, excavation, and archaeobotanical investigation case for offer an important comparative slightly earlier Byzantine of the riparian districts around Melitene and Samosata occupation in the tenth and eleventh centuries.

centers with their grain markets that craved mainly wheat and barley, a natural fallback would have been rye,which competed poorly in the urban market, but is an admirable crop in regional and local economies of necessity.74

In the fertile Euphrates valley around Gritille, middle-Byzantine-era farmers grew a of melange but wheat and barley dominated, with wheat crops, responsible for of the crops produced there. Limited quantities of grapes, figs, cotton, flax, and nut crops suggest theminor role of fiber as beans and vetch, were and fruit plants. Leguminous plants, such 33-68 percent a grain, these plants suggest relatively well-managed agrarian cycle, with legumes fixing nitrogen was also in the soil for the grain crop that followed. Rice probably and, like cotton, required irrigation. Basalt oil-mill compo grown, also produced; grown in rotation with

screw nents and possibly also weights suggest that the region pro duced some form of oil crops, perhaps sesame, but it is not impossible that olives were grown. As elsewhere

and in the Konya plain, the along the Euphrates sherd scatters on the river terraces around Gritille attest that the inhabitants, like the late antique farmers of the region, their fields as part of a high-input, managed farming regi a men. Manure was in part of livestock. Pigs diversity by provided was a were themost common animal, and in the key ingredient pork medieval manured diet, one clear indicator around Gritille eastern Anatolia, that the late medieval predominantly the core settle sought provender outside of over wooded ments to which and grassy they belonged and foraged rear acorns and other fibrous crops. This mode of areas, eating hog a form of swine herd represents wide-ranging ing, called pannage, areas of to landscapes where large ing, nonintensive and best adapted pigs untilled close to villages. also were important to the late medieval popu Sheep and goats As with swine production, caprines were part of a lation of Gritille. subsistence economy rather than a more specialized economic niche, on local such as wool or dairy production. Herds were maintained resources, not sent far afield in semi-nomadic grazing forays, and land are available the aims of their herders were meat and a steady supply of provide Farmers kept large cattle sufficient levels of breeding stock. to remained in and people living In this part of Christian.

maintain

"Late 74 (Beycesultan) H. Haelbek, Bronze Age and Early Byzantine Crops at Beycesultan," AnatSt n (1961): 77-97.

262

MICHAEL

DECKER

(bovines) primarilyfor traction;dairy productionwas a secondary

concern,
two

and meat was

apparently

simply

byproduct

of the latter

concerns.75

activities. Fairly small herd sizes are implied, with produc a tion aiming tomaintain, not pri enlarge these herds. This suggests mary crop focus of heavily worked fields, a fact corroborated by the nomadic of the river terraces. indicating widespread manuring In turn, the major focus on cropping suggests substantial numbers of settlements and a landscape that rendered necessary such agropas sherd scatters

removed from the villages) and rough country was the primary form of herding, rather than transhumant or other mid-long-distance

economy from Gritille reflects a land in several scape that I believe parallels that of the Anatolian plateau use of outfields (lands more than i km respects. First, the generally The late medieval herding

toral strategies, where animals were an important part of a system whose focus remained grain and leguminous crops. Environmental studies at Amorium have yielded finds of culti

vated barley and wheat as well as millet.76 Had Leo of Synada traveled to Amorium, his the fifty or so miles from Synada hunger forwheat a cup of bread would have been sated. He also could have enjoyed several screw-press elements have been found in the city, and these argue for surplus that would have been traded at least the local wine:

more of medi locally. Even interesting is the presence at Amorium a that may indicate a system of known as ley cago, legume production or convertible husbandry. Ley farming enabled the produc farming tion of cereal crops and animals in an efficient and intensive rotation. In this rotation cereal crops alternated with inwhich long-fallowing, the fields were with two pur The planted medicago. legume served poses: first, to graze animals, and second, to fix precious nitrogen in the soil to restore it. While ley farming is known from the Roman period cation late antiquity via the evidence for its appli Geoponika, inMiddle Anatolia had been Byzantine missing. This small detail provides a a powerful reminder of how seemingly minor piece of data can significantly enhance our archaeological knowledge. and

Conclusions
in this study of two frontiers. The first is that along the corridor where Byzantium met Persia until the sev Tigris-Euphrates enth century. Byzantine political control in this zone collapsed in the mid-seventh century. In and of itself, this change in political masters need not have entailed any alteration of the status quo; the farmers on the not have been in a number of ground need displaced. However, border areas, the or suffered farming population apparently relocated severe losses. The second frontier centered on touched Cappadocia, I have written
75 G. Stein, "Medieval Pastoral inRedford, Production Systems at Gritille,"

The Archaeology (above, n. 19). 76

of theFrontier, 181-209 and Neil Christie, 1992 Interim

R. M. Harrison

"Excavations

at Amorium:

Report," AnatSt 43 (1993): 124-25; C. S. Lightfoot, E. A. Ivison, et al., "Amorium Excavations 1994: The Seventh Preliminary 45 (1995): 152-53. Report," AnatSt

BYZANTINE

EAST

FRONTIER

ECONOMY

263

Levant. The eastern landscape of this frontier from Cilicia to Melitene a fractious and uneven endured history. At present, the evidence from and limited excavation suggests that the popu survey archaeological sometime in the centuries. The steeply seventh/eighth in site numbers around Samosata was precipitous. Most fall of the lation declined small outlying sites, evidence of the spread of the early-Byzantine-era over the return was retarded landscape, disappeared. Their population centuries by warfare and those for political uncertainty. Whether

the Euphrates,

and trailed down

to the coastlands

of the northeastern

succumbed to pestilence, war, or other natural disasters populations is at the moment the subject of mere speculation. But the transition was not without trauma. Given the reaction of the toMuslim rule communities of northern Mesopotamia, expressed Aramaic-speaking in the Muslim of pseudo-Methodius, Apocalypse political sharply control visited Christian communities with sustained and memorable One can only expect that the endemic warfare engendered hardship.77 on the exposed by the annual jihad and Byzantine raids wrought havoc towns and country of the upper northern Syria, and Euphrates valley, Cilicia. cated, resources dwindled,

or were relo In the face of this danger, emigrated populations were troubled, and the communications

of the formerly densely packed countryside shrank. population For nearly the whole of our period, the Taurus and Anti-Taurus Mountains frequently divided was the and Muslims, both Christians transgressed by an arena of interaction two civilizations, peaceful meeting ground of where ideas and bloodlines were exchanged, and a zone of violence the Byzantine from theMuslim worlds. This line,

where border lords carved out reputations and estates. The pattern of settlement, farmore difficult to discern than that farther east, shared certain features with tices evidenced characterized marked into the less desirable in traces of manured the Euphrates valley. The intensive farming prac fields, the same encroachment

cities were never as large or as ubiquitous Prima and Secunda were, on the In Justinian sday Cappadocia plateau. in terms of conventional Graeco-Roman cities, underdeveloped. differences as well: In late antiquity,

areas, and the creation of agricultural outliers late antique settlement on the plateau. But there were

of the of the eastern borderlands large swathes more in any other were settled densely than Empire Byzantine recorded period. In Cilicia, North Syria, and along the Euphrates, the to convert to landscapes sought underexploited growing population soils. As they pushed and everywhere invaded marginal agriculture, into the farther limits of the arable, farmers worked ever more inten tomaintain a range of soil fertility and strategies sively and employed sustain their livelihoods. Settlement along the Euphrates, as in thereby
77 P-Alexander, The Byzantine (Berkeley, 1985), 36-51. Apocalyptic Tradition

264

MICHAEL

DECKER

was both expansiveand dense,with villages, Cilicia andNorth Syria,

estates, and humbler farmsteads

were stall feeding or folding was desirable. increasingly restricted, and to avoid Animals had to be restricted in theirmovements, damaging over stubble in folds or stalls, or close Confinement crops. herding

but rangelands Animals fitted into thishigh-inputagriculture, side.

spread

over a

deeply

worked

country

did not inevitably lead to conflict. The complex relation in the ships that developed between nomad and settled populations a range cannot be pre-Islamic world explored here, but they included of action and reaction. Farmers tapped into the available labor force conditions offered by nomadic groups during the harvest, and no doubt wel comed the vast flocks into grain lands where dung was desperately needed tomaintain fertility.As they had done for centuries, nomads

after the grain crops were taken, rendered fertilizer accessible. In the Jazira and elsewhere in Byzantine Syria, transhumants far and wide, but they saw their grazing lands encroached ranged upon by sedentaries who pushed into the traditional grazing lands in nomadic pasturelands dimin search of agricultural land. Although such pressed ished because of the increase in the farming population,

At exchanged pastoral products for grain and other farm products. other times, they raided, rustled, and kidnapped from the farms and of the frontier. In Byzantine Anatolia, the flocks of sheep and villages herds of horses land magnates nomadic that represented the mobile capital of the powerful dominated. Before the arrival of the Turks, the purely

none of the estate records of the pastoralists. We have we know that, since highland magnates, but antiquity, Cappadocia was horse need vast areas of quality pasture and country. Horses access to water. The pressure exert on the ready they countryside is forwandering border lords had both substantial, their value high. The Cappadocian the incentive and themeans to tightly control their domains, and it is

component of the population was severely restricted: the estates of the elite, with their intensive farm expansive Byzantine herds, would have allowed little room ing practices and managed

sort ofmutualism thatwe saw in operation certainly possible that the in Syria and to the presence of was, due Mesopotamia large, centrally estates, simply not possible inAnatolia. While agropastoral managed we should envision movements of herds from summer large seasonal to winter pastures, most of these animals men, but to estate owners like Philaretos. To conclude, belonged not to the herds

thiswork has dealt with major interlocking elements of life in the eastern frontiers of Byzantium from the seventh to the eleventh centuries: the nature of settlement based on the current data and the character of archaeological agricultural life in the fron tier zones. The present data, and incomplete fragmentary though they

BYZANTINE

EAST

FRONTIER

ECONOMY

265

are, suggest that the Islamic conquests did mark awatershed moment within the settlement history of Byzantium. The of the populations and the northeastern corner of the Levant fell from Euphrates valley

late antique levels that they have in our day. Along the regained only new frontier of that bordered theMuslim lands, Cilicia-Cappadocia the old late antique agricultural regime apparently sisted throughout the middle Byzantine period. survived and per

The

to say of the way that the east something also ern Anatolian was farmed and controlled, and how the landscape inhabitants maintained and expanded their mas Byzantine-period tery over the region. It is somewhat ironic that the difficult landscape of the plateau that defined the rural world was, in many ways, con evidence has ducive to the defensive raids. Byzantine amobile such a regime was in place centuries before capital economy: the coming of Islam. Nor did the Byzantine farmers of the medi were scarce. At eval period abandon their lands. Arable lands always those in the face of theMuslim agriculture needed to create Anatolia did require a shift to pastoralism

one where soil and water infrequent points good converged, found settlement as well. These farming oases were both compact and dispersed. This fact was both a blessing and a curse: there was structures too little arable land from which to develop large urban investment in hydraulic engineering, but the without overwhelming smaller pockets of land lent themselves to defense and the integration of large numbers of animals, whose mobility could take advantage arable has always been of the large, intervening empty spaces. Good no less so in medieval As the popula in short Cappadocia. supply, to in the ninth and tenth centuries and Byzantine tion expand began confidence recovered, the frontier aristocracy competed intensely

the landed estates that formed the only and assiduously developed sure kernel whence their power could germinate. In this process, the from it. By weaken invasions were not the death knell?far Muslim in the regions, and by provincial ing the power of Constantinople the emperor to give up most, ifnot all, of the imperial estates forcing the local lords were free towax that had once blanketed Cappadocia, tree

to the fullest. Only centuries laterdo we find the full fruitsof the
and Bardas Phokas, in ripen: in the rebellions of Bardas Skleros The power of these fami the ostentation of Eustathios Maleinos. a substructure of on a superstructure of violence lies rested cladding

new frontier. along the

wealth thathad been built block by block in farmsand fields landed

University SouthFlorida of

266

MICHAEL

DECKER

Acknowledgments
I received

Iwould like to thank Dumbarton Oaks and theparticipantsat the 2005 Spring Symposiumfor their hospitalityand for the comments
there. I would also as well on this piece, like to thank Alice-Mary for Talbot as thank the reviewers from whose Florida

her work

comments I benefitedgreatly. Nick J. Finally, Iwould like to thank

Maroulis, have made whose contributions possible much to the University of South of the research for this article.

BYZANTINE

EAST

FRONTIER

ECONOMY

267