Content uploaded by Ido Shahar
Author content
All content in this area was uploaded by Ido Shahar on Jan 21, 2018
Content may be subject to copyright.
© Koninklijke Brill NV, Leiden, 2012 DOI: 10.1163/156852012X628482
Journal of the Economic and
Social History of the Orient 55 (2012) 1-31 brill.nl/jesh
Irrigation in the Medieval Islamic Fayyum:
Local Control in a Large-Scale Hydraulic System
Yossef Rapoport and Ido Shahar*
Abstract
Because of the unique set of sources available, the Fayyum in Middle Egypt oers a unique
case study of large-scale irrigation from antiquity to the Islamic period. A close reading of
a cadastral survey of the province from 641/1243-4 shows that the distinctive aspect of the
Islamic period was the local control of water supply and management. Drawing on the
engineering experience of the villagers, water allocation and management in the gravity-fed
canals of the Fayyum were in the hands of iqṭāʿ holders and tribal groups along the main
canals, a pattern similar to that which pertained in mediaeval al-Andalus.
Grâce à une série de sources exceptionnelles, le vaste système d’irrigation du Fayoum, en
Moyenne-Égypte, peut être reconstitué depuis l’Antiquité jusqu’à l’ère islamique. L’examen
approfondi d’un relevé cadastral de cette province, datant de 641/1243-4, montre que
l’époque islamique se caractérise par une gestion locale de l’approvisionnement en eau.
S’appuyant sur l’expérience technique des villageois, les détenteurs d’iqṭāʿs et les groupes
tribaux implantés le long des principaux canaux du Fayoum contrôlaient la répartition de
l’eau et la gestion des canaux alimentés par gravitation. Ce modèle de gestion de l’eau rap-
pelle à bien des égards les pratiques en usage au Moyen Âge en al-Andalus.
Keywords
Fayyum, Egypt, Nile, irrigation, iqṭāʿ, al-Nābulusī, al-Andalus, al-Lāhūn, tribes
Few places on earth have been as dependent on irrigation works as the region
of the Fayyum, in Middle Egypt. e Fayyum lies in a large depression of
*) Yossef Rapoport, School of History, Queen Mary University of London, Mile End Road,
London E1 4NS, UK, y.rapoport@qmul.ac.uk; Ido Shahar, Department of Middle East
Studies, Ben Gurion University of the Negev, P.O. Box 653, Beer Sheva, 84105, Israel. In
the preparation of this paper, we have beneted from generous input by Brendan Haug and
Dominic Rathbone. David Price has kindly provided us with an electronic copy of his PhD
dissertation.
2 Y. Rapoport, I. Shahar / JESHO 55 (2012) 1-31
the Libyan Desert, 25 kilometres west of the Nile, to which it is linked by
a long canal known as the Baḥr Yūsuf. Because of its topography and loca-
tion, agriculture in the Fayyum relied entirely on a system of embank-
ments and canals that allowed the water of the Nile to reach the depression
in sucient quantities, but without ooding the cultivated areas. e agri-
cultural exploitation of the Fayyum began in the Middle Kingdom, when
the pharaohs of the Twelfth Dynasty transformed the Fayyum into a giant
holding basin for excess water during the annual Nile oods and a reser-
voir of Nile water for irrigation. Under the Ptolemaic dynasty, in the fourth
and third centuries BCE, the reservoir was drained, and a large-scale recla-
mation project tripled the area of cultivated land.
e principal elements of the Ptolemaic irrigation system were an
embankment at the entrance to the depression to regulate the amount of
water entering it, the construction of main and subsidiary irrigation sup-
ply canals, and the drainage of excess water into Lake Qarun, in the north-
ern part of the depression. Once drained and put into cultivation, the
Fayyum became one of the most fertile and prosperous provinces of Egypt,
a position it has maintained up to the present.1
In this millennia-long history of irrigation, the mediaeval Islamic period
is generally perceived as a period in which the Ptolemaic achievements
were eroded. A process of decline is seen as setting in during the third and
fourth centuries CE, when many villages around the edge of the Fayyum
ceased to receive adequate water supplies and were partially or wholly
abandoned. More than a century of excavations of sites in the northeast-
ern, northwestern, and southern edges of the Fayyum unearthed a series of
large and rich settlements, which were, by and large, deserted by the time
of the Arab conquest. e desertion of these sites, a result of the failure of
the canals running along the foot of the mountains surrounding the prov-
ince, is often viewed as part of a general decline. e shrinking of the cul-
1) For recent accessible summaries of the history of the Fayyum in antiquity, see Roger S.
Bagnall and Dominic W. Rathbone, Egypt: From Alexander to the Copts: An Archaeological
and Historical Guide (London: British Museum Press, 2004): 157.; R. Neil Hewison, e
Fayoum: History and Guide (Cairo: e American University of Cairo Press, 2001); Willy
Clarysse, “e Fayum: A First Introduction,” in e Fayum Project (http://www.trismegistos
.org/fayum/index.php) [accessed 22 July 2011]. For a survey based on pollen records, see
P. Mehringer et al., “A Pollen Record from Birket Qarun and the Recent History of the
Fayum,” Quaternary Research 11 (1979): 238-56.
Irrigation in the Medieval Islamic Fayyum 3
tivable area persisted through the subsequent Islamic period, until irrigation
was revived in the nineteenth century.2
As evidence for decline during the Islamic period, scholars have relied
on the unique tax register of Abū ʿUthmān al-Nābulusī, an Ayyubid o-
cial who visited the province during the winter of 641/1244-5. e work
has been known to scholars since the end of the nineteenth century, when
it was published under the title Taʾrīkh al-Fayyūm wa-bilādihi (“History of
the Fayyum and its Villages”).3 As noted by John Ball, in his classic study
of the history of irrigation in the Fayyum, al-Nābulusī formulated his work
as an attempt to revive agriculture in the Fayyum, whose aairs, he claimed,
had deteriorated through negligence. One of the chapters of the work is, in
fact, entitled “An account of the deterioration (taghayyur) of its canal and
the reason for that, and of the villages that have so fallen into ruin that
their reconstruction could be achieved only by [the investment of ] gener-
ous sums of money over a long period of time.” Al-Nābulusī specically
reports that the central government in Cairo recorded no expenditure on the
irrigation system in the Fayyum for more than a century.4
Not all assessments of irrigation as reected by al-Nābulusī’s survey have
been so negative. Ali Shafei Bey, the irrigation inspector of the Fayyum
during the 1930s and 1940s and an amateur historian, was impressed with
the achievements of mediaeval irrigation, particularly with the amount of
land under extensive perennial cultivation. For him, under the Ayyubids
the “Fayoum Province had an irrigation system very similar to, if not bet-
ter, than present.”5 More recently, Sato’s work on mediaeval rural society
highlights the extent of investment in irrigation under the Mamluk sul-
tans, especially in the seventh/thirteenth and eighth/fourteenth centuries.
His work uses details of irrigation works from al-Nābulusī’s survey to illus-
trate the complexity and dynamic nature of the management of irrigation
2) is view is most clearly articulated in A. E. R. Boak, “Irrigation and Population in the
Faiyum, the Garden of Egypt,” e Geographical Review 16/3 (1926): 353-64. is is also
the view of John Ball, Contributions to the Geography of Egypt (Cairo [Bulaq]: Government
Press, Survey and Mines Dept, 1939), 199-220; and, more recently, David H. Price, “e
Evolution of Irrigation in Egypt’s Fayoum Oasis: State, Village and Conveyance Loss,” PhD
diss., University of Florida, 1993.
3) Abū ʿUthmān al-Nābulusī, Taʾrīkh al-Fayyum wa-bilādihi, ed. B. Moritz (Cairo: Publica-
tions de la Bibliothèque Khédiviale, 1898).
4) Ball, Contributions: 220; see also Price, “Evolution,” 185-94, al-Nābulusī, Taʾrīkh: 6.
5) Ali Shafei, “Fayoum Irrigation as Described by Nabulsi in 1245 A.D.,” Bulletin de la Société
Géographique Royal d’Égypte 20 (1940): 283-327, at 285.
4 Y. Rapoport, I. Shahar / JESHO 55 (2012) 1-31
in mediaeval Egypt.6 Sato’s conclusions are accepted by Stuart Borsch, who
views the rule of the Ayyubid and early Mamluk sultans as a zenith in
investment in irrigation, before a subsequent collapse brought about by
the Black Death.7
In recent years, impetus to revisiting the history of irrigation in Islamic
Fayyum and al-Nābulusī’s work in particular has come from scholars of
late antiquity. In trying to explain the desertion of sites on the edges of the
province in the third and fourth centuries CE, these scholars prefer to see
a process of adjustment rather than a failure of irrigation works. In a survey
of the south-western edges of the Fayyum, Kirby and Rathbone note the
great variability in the history of the sites, suggesting that the villages in the
area were deserted gradually, over many centuries. Such a pattern does not
t the traditional account of a sudden catastrophic collapse of the irriga-
tion system.8 Keenan goes further: he points out that all the deserted vil-
lages are found on the edges of the province, while we know little about
the history of Greco-Roman settlements in the centre of the depression.
Evidence from a handful of deserted villages located on the outskirts may
thus not be typical of the province as a whole.9 Keenan, in particular, has
called for a sustained study of al-Nābulusī’s rich text in order to gain a
fuller understanding of the development of the Fayyūm and its water
regime.10
is essay aims to provide a systematic and critical account of the irriga-
tion system of the Fayyum in the seventh/thirteenth century, as described
by al-Nābulusī. Rather than taking al-Nābulusī’s various statements in iso-
lation, the aim here is to explore the totality of the rich text—the most
6) Tsugikata Sato, State and Rural Society in Medieval Islam (Leiden: Brill, 1997): 227-8.
7) S. Borsch, “Environment and Population: e Collapse of Large Irrigation Systems
Reconsidered,” Comparative Studies in Society & History, 46/3 (2004): 451-68; e
Black Death in Egypt and England: A Comparative Study (Austin: University of Texas Press,
2005): 38.
8) C. Kirby and D. Rathbone, “Kom Talit: e Rise and Fall of a Greek Town in the Fay-
ium,” Egyptian Archeology 8 (1996): 29-31.
9) J. Keenan “Deserted Villages: From the Ancient to the Medieval Fayyum,” Bulletin of
American Studies in Papyrology 40 (2003): 119-40; See also B. Kraemer, “e Meandering
Identity of a Fayum Canal: e Henet of Moeris / Dioryx Kleonos / Bahr Wardan / Abdul
Wahbi,” American Studies in Papyrology (2010): 365-76.
10) J. Keenan, “Fayyum Agriculture at the End of the Ayyubid Era: Nabulsi’s Survey,” in
Alan K. Bowman and Eugene Rogan (eds.), Agriculture in Egypt from Pharaonic to Modern
Times (Oxford: Oxford University Press, 1999): 287-99; J. Keenan, “Landscape and Mem-
ory: Al-Nabulsi’s Taʾrikh al-Fayyum,” Bulletin of the American Society of Papyrologists 42
(2005): 203-12.
Irrigation in the Medieval Islamic Fayyum 5
detailed cadastral survey to have survived for any region of the mediaeval
Islamic world—in its historical context, using both qualitative and quan-
titative research methods. is work is based on a close reading of the
treatise and on an analysis of the terminology and scal data provided by
al-Nābulusī. It is also based on the scal data contained in the work, which
is now publicly available on a dedicated website.11 We complement
al-Nābulusī’s survey with a fth/eleventh-century account of the Fayyum
irrigation system, written by a certain Jaʿfar Abū Isḥāq and preserved by
al-Maqrīzī (d. 845/1442).12 Another account of the irrigation of the
Fayyum in the mediaeval period, found in the administrative manual of
Ibn Mammātī (d. 606/1209), relies largely on that of Abū Isḥāq.13 Refer-
ences to the Fayyum in the works of earlier Muslim scholars, such as
Ibn ʿAbd al-Ḥakam (d. 257/871) and al-Masʿūdī (d. 345/956), are espe-
cially useful for their glimpses of the transition from late antiquity to the
Islamic era.
is study of the irrigation of the Fayyum in the Islamic period relates
to two larger questions that have dominated the study of irrigation in the
mediaeval Islamic period. One question has been whether irrigation sys-
tems improved or declined under Islamic rule. Andrew Watson has viewed
the transfer of irrigation technologies from Iran and India as part of the
“mediaeval green revolution” brought about by the spread of Islam, and, in
detailed studies of irrigation in al-Andalus, it has been shown that Muslims
greatly extended the network of irrigation canals they found in place.14
Peter Christiansen, on the other hand, has chronicled the decline of large-
scale irrigation in Iran, while other archaeological studies have found that
Islam brought no change at all or failed to stem the late-antique collapse of
irrigation infrastructure in the Near East.15
11) Y. Rapoport and I. Shahar, “Rural Society in Medieval Islam,” (http://www.history
.qmul.ac.uk/ruralsocietyislam) [accessed 22 July 2011].
12) Al-Maqrīzī, al-Mawāʿiẓ wa-l-iʿtibār fī dhikr al-khiṭaṭ wa-l-āthār, ed. A. F. Sayyid (Lon-
don: Muʾassasat al-Furqān li-l-Turāth al-Islāmī, 2002-4), 1:669-74.
13) Ibn Mammātī, Qawānīn al-dawāwīn (Cairo: al-Jamʿiyya al-Zirāʿiyya al-Malikiyya,
1943): 229-32; translation in R. S. Cooper, “Ibn Mammati’s Rules for the Ministries,”
PhD diss., University of California, Berkeley, 1973): 74-7.
14) T. Glick, “Hydraulic Technology in al-Andalus,” in e Legacy of Muslim Spain, ed.
S. K. Jayyusi (Leiden: Brill, 1992): 974-86.
15) P. Christiansen, e Decline of Iranshahr: Irrigation and Environment in the History of
the Middle East 500 B.C. to A.D. 1500 (Copenhagen: Museum Tusculanum Press, 1993);
M. Decker, “Plants and Progress: Rethinking the Islamic Agricultural Revolution,” Journal
of World History 20/2 (2009): 187-206.
6 Y. Rapoport, I. Shahar / JESHO 55 (2012) 1-31
A second question has been the degree of centralization required to con-
trol large-scale irrigation systems. Egypt in particular has been the focus of
the debate over Wittfogel’s Oriental Despotism, in which a link has been
posited between strong state structures and the demands of centralized
control of the Nile annual ooding.16 In contrast, a dominant theme of the
study of rural society in mediaeval al-Andalus has been the predominance
of local control. omas Glick, building on previous studies by Barcelo,
Guichard, and Bazzana, argues that, in almost all irrigation systems, except
for the large-scale macro-systems, tribal organization has been particularly
suitable for managing rights to water and resolving conicts between
upstream and downstream communities.17
In relation to these two questions of continuity and centralization, and
in light of the unique topography of the Fayyum, this paper looks at the
methods of water supply, allocation, and management in mediaeval Islamic
Fayyum. It argues for a signicant degree of continuity with pre-Islamic
infrastructure, yet also shows that, at least up to the middle of the seventh/
thirteenth century, the trend has been one of decentralization and localiza-
tion of knowledge and control. With practically no irrigation bureaucracy,
few direct irrigation taxes, and minimal direct interference, the manage-
ment of the irrigation system appears to have been very much in the hands
of local communities, which were, as in al-Andalus, organized in tribal
groups. e decentralization and localization of control and knowledge
may have limited the scope of large-scale irrigation projects that required
heavy investment and professional engineers, but decentralization did not
necessarily mean decline. Mediaeval Islamic Fayyum had a fully function-
ing irrigation system, one which supported a thriving economy and which
continued to develop.
e al-Lāhūn Dam
e dam at al-Lāhūn is an enormous earthen dyke that can still be seen,
from al-Lāhūn to the mortuary pyramid at the entrance to the Fayyum. It
16) K. Wittfogel, Oriental Despotism: A Comparative Study of Total Power (New Haven: Yale
University, 1957).
17) T. Glick, From Muslim Fortress to Christian Castle: Social and Cultural Change in Medi-
eval Spain (Manchester: Manchester University Press, 1995): 69-76; D. M. Varisco, “Sayl
and Ghayl: e Ecology of Water Allocation in Yemen,” Human Ecology 11 (1983), 365-83
(esp. 378-9), reprinted in his Medieval Folk Astronomy and Agriculture in Arabia and the
Yemen (London: Ashgate, 1997).
Irrigation in the Medieval Islamic Fayyum 7
awaits an archaeological study, but its relation to the pyramid suggests it
was rst constructed in the Middle Kingdom to divert the Bahr Yūsuf into
the depression.18 ere are several descriptions of the al-Lāhūn dam in
geographical texts from the Islamic period, and they all assert that the pur-
pose of the dam was to divert the water of the Baḥr Yūsuf into the Fayyum,
while allowing high Nile oods to escape back to the Nile. Al-Masʿūdī, the
rst Muslim scholar to give a technical account of the operation of the
al-Lāhūn dam, reports that the biblical-Qur’anic Yūsuf ( Joseph) built the
barrage at al-Lāhūn in order to allow the appropriate amount of water to
enter the depression. Openings within weirs (qanāṭir) allowed the water to
go through them rather than over the dam.19 e next account of the bar-
rage, by al-Muqaddasī (. 375/985), suggests that the weir mechanism has
gone out of service, to be replaced by a simpler method of regulation.
Al-Muqaddasī describes a simple spillway dam, in which water goes over
the crest of the embankment when the Nile waters are high, which also
allows boats to sail in and out of the Fayyum, albeit with diculty. At the
bottom of the barrage there were glass pipes, which served as outlets
(manās). When the Fayyum has received enough water, these pipes are
opened, so that the level of the water behind the barrage drops.20
e most detailed technical account of the dam in the Islamic period,
by Abū Iṣḥāq in 422/1031, conrms the abandonment of the weirs and
sluice-gate system, which was still visible.21 e area in front of this section
was paved, and in the embankment one can still see ten ancient stone weirs
(qanāṭir), with sluice gates (abwāb). ese pre-Islamic weirs at the end of
the embankment formerly diverted the water to the Fayyum through an
old canal, which, he says, was no longer used. Like al-Muqaddasī, he men-
tions glass outlets (manās) in the embankment, through which water
escapes.
e level of water entering the Fayyum was controlled by two openings,
which allowed excess water to cross the embankment when the Nile waters
were high. Abū Iṣḥāq describes in detail a large opening, in the southern
section of the embankment, which was 120 cubits wide. e two banks of
18) Bagnall and Rathbone, Egypt: 152; cf. an earlier view in Ball, Contributions, 213-22.
19) Al-Masʿūdī, Murūj al-dhahab wa-maʿādin al-jawhar, ed. Ch. Pellat (Beirut: al-Jāmiʿa
al-Lubnāniyya, 1965-79), 2:72 [no. 784], 2:80 [nos. 797-8].
20) Al-Muqaddasī, Kitāb aḥsan al-taqāsīm fī maʿrifat al-āqālīm (Leiden: Brill, 1906): 208.
21) Al-Maqrīzī, al-Mawāʿiẓ, 1:670-1; see also Shafei, “Lake Moeris and Lahun Mi-wer and
Ro-hun: e Great Nile Control Project Executed by the Ancient Egyptians,” Societé
d’Égypte Bulletin 33 (1960): 187-215, esp. 210-3.
8 Y. Rapoport, I. Shahar / JESHO 55 (2012) 1-31
this opening sloped towards each other until they reached a depth of four
cubits, which allowed boats to pass through it during the Nile ood. At
other times of the year this opening was closed by a dyke (of grass) called
Lamsh, which was forty cubits wide and covered the lowest past of the
opening. A smaller opening, in the northern section of the embankment,
was twenty cubits wide and two cubits deep. is opening is also blocked
by a dyke of grass, called Lknd.
By al-Nābulusī’s time, in the seventh/thirteenth century, the memory of
the ancient pre-Islamic structure at al-Lāhūn had become blurred. He no
longer mentions sluice gates or weirs, which suggests that these structures
were no longer visible. He does, however, assert, on the authority of local
“wise men,” the existence of culverts and paving in the section of the
embankment closest to al-Lāhūn; this ancient apparatus, which prevented
the silting of the canal, were, by his time, obsolete.
Like Abū Iṣḥāq, al-Nābulusī describes in detail an opening in the embank-
ment that allowed boats to sail into and out of the Fayyum during high Nile
oods. is must be the same four-cubit-deep opening in the southern sec-
tion of the embankment that is called lamsh by Abū Iṣḥāq. Al-Nābulusī says
that water escapes through this opening during the Nile oods, and boats
pass through it, as they do not want to risk passing over the dam itself.22
Given that the sluice gates were no longer used, the main annual mainte-
nance work on the al-Lāhūn dam was the annual blocking of the shipping
opening in the embankment. When the Nile ood had peaked and begun to
recede, it was necessary to block the opening in order to divert into the
Fayyum whatever water still owed in the Baḥr Yūsuf. Al-Nābulusī provides
a lively account of this process, peppered with uncharacteristic hints of Egyp-
tian dialect:
when the Nile recedes . . . the piece (qiṭʿa) is installed at al-Lāhūn. . . . e “piece” is a
long palm log to which straw and rags are xed. ese are tied up with ropes, so that
it becomes very thick. e strong ropes are at its edge, and the ends of the ropes are in
the hands of a large group of men on the bank adjacent to the small village (ḍayʿa)
called al-Lāhūn, and on the opposite bank. ey release the ropes little by little, while
the water carries the piece and pulls it toward the opening . . . releasing it little by little,
until it comes to the mouth of the opening and blocks it and thereby prevents the
water from escaping. en the men pile soil and clay on it so that it resembles the bank
22) e second, smaller opening to the north, which is called lknd by Abū Iṣḥāq, is not men-
tioned by al-Nābulusī (Taʾrīkh: 12). He does, however, mention the term lknd as a name for
the barrage as a whole, or a section of it (Taʾrīkh: 101).
Irrigation in the Medieval Islamic Fayyum 9
adjacent to the structure, so much so that a person may cross over the dam from
al-Lāhūn to the bank of Qāy,23 just as he would proceed on the same bank.24
e account exemplies clearly the way the upkeep and regulation of the
al-Lāhūn dam had changed in the Islamic period. e process was simple,
and based on local knowledge and materials. Moreover, although the blocking
of the opening in the embankment was essential for the water supply of the
Fayyum as a whole, there is no evidence of hierarchical or centralized control.
No state ocial is mentioned, nor any order issued by the governor of the
province, nor corvée labour; rather, al-Nābulusī simply describes “a large
group of men from the villages of al-Fayyum, as well as engineers, [who]
gather together.”25 e local, simple, decentralized upkeep of the dam kept
it functioning, and the irrigation system of the Fayyum did not collapse.
Yet, the local and decentralized maintenance work on the al-Lāhūn dam
corresponds with the way the dam itself has changed in the Islamic period,
as the sluice gates and weirs, whose maintenance must have been labour-
intensive, were abandoned. Perhaps for that reason, during the seventh/
thirteenth century the Fayyum did face an increasing problem of water
supply. e local and government responses to these problems provide an
excellent case-study of the management of the Fayyum irrigation system.
Water Shortage and Local Investment
e key problem facing the agriculture of seventh/thirteenth-century Fayyum
was a shortage of water, a shortage caused partly by the decreasing ow
from the Nile, through the head of the Baḥr Yūsuf canal, 300 kilometres
south of the Fayyum. In the past, says al-Nābulusī, the head of Baḥr Yūsuf
had owing water for eight months of a year and was dry for only four
months. In his time, however, the situation is reversed, and water ows
into it only four months in a year, and it is dry during the remaining eight
months. e Fayyum still contained some water in the other eight months
of the year, because seepage from underground sources (nabʿ ) continued
to feed the canal after the Nile inundation ended. But, without a constant
ow from the Nile, water became scarce. Moreover, many waterwheels on
the banks of the Baḥr Yūsuf in the upstream provinces of al-Ashmūnayn
23) Qāy is a village located several miles southwest of al-Lāhūn.
24) Al-Nābulusī, Taʾrīkh: 12; see also Shafei, “Fayoum,” 307.
25) Al-Nābulusī, Taʾrīkh: 12.
10 Y. Rapoport, I. Shahar / JESHO 55 (2012) 1-31
and al-Bahnasā drew o water for irrigation, at the expense of the Fayyum.
Al-Nābulusī suggests closing o these upstream waterwheels when the
Nile recedes.26
According to al-Nābulusī, the water shortage in the Fayyum was not a
result only of diminished supply from the Nile but also of problems at the
al-Lāhūn dam itself, where silting caused the shoaling of the canal.
Al-Nābulusī attributes the silting of the canal to the blocking of the ancient,
pre-Islamic culverts, which allowed silt to pass through. Whether this was
indeed the cause of the silting is impossible to say. At any rate, he reports
that soil and clay had accumulated in front of the embankment, so that the
height (irtifāʿ ) of the al-Lāhūn embankment above the bed of the canal was
reduced from fteen cubits to seven cubits or less. is meant that more
water escaped back to the Nile, and less water was available for agriculture
in the Fayyum.27
Al-Nābulusī links the water shortage in the Fayyum to the abandon-
ment of villages at the edges of the province, along the two major canals
that branch out soon after the water enters the Fayyum.28 Al-Nābulusī is
describing here the culmination of a long-term process that started, as we
know from excavations, by the third century CE, or even earlier. Abū Iṣḥāq,
writing in the fth/eleventh century, also notes that there are abandoned vil-
lages along the Tanabt ̣awiya canal and the easternmost canal, which he calls
the al-Awāsī canal.29 Although the immediate reason given for the desertion
of the sites is lack of canal maintenance, it seems clear that these canals,
which followed a course higher up along the base of the mountain, would be
the rst to suer a water shortage.
Another indirect evidence of the shortage of water in the Fayyum is the
relative insignicance of the reservoir, known as al-Gharaq, in the south-
western corner of the depression. e purpose of this reservoir was to store
excess water owing into the depression during the inundation. In pre-
Islamic times it had been an important source of water during the drier sea-
sons, and it is referred to repeatedly in Ottoman sources as a massive body of
water.30 In al-Nābulusī’s work, however, the references to the al-Gharaq
reservoir are few. Only three minor villages in western Fayyum—Muqrān,
26) Al- Nābulusī, Ta ʾrīkh: 11-12.
27) Al-Nābulusī, Taʾrīkh: 15.
28) Ibid.: 17-18.
29) Al-Maqrīzī, al-Mawāʿiẓ, 1:671.
30) For the Ottoman period, see A. Mikhail “An Irrigated Empire: e View from Ottoman
Fayyum,” International Journal of Middle East Studies 42: 569-90. On the al-Gharaq reser-
Irrigation in the Medieval Islamic Fayyum 11
Diqlawa, and Masjid ʿĀʾisha—were irrigated by water from al-Gharaq, by
two dierent outlets (one for Muqrān, and another for Diqlawa and Masjid
ʿĀʾisha). e three villages irrigated from the reservoir were marginal to
the Fayyum economy; their combined tax revenue amounts to less than
1 percent of the total tax revenue of the province.31 It seems, therefore, that
the reservoir was not fully functional and that only a small part of it retained
water, further suggesting a low level of water supply to the depression.32
While al-Nābulusī blames this water shortage on lack of investment by
the central government, he also leaves us a fascinating account of expensive
collaborative attempts by the provincial iqṭāʿ (land-grant) holder and local
villagers to increase the water supply to the Fayyum, attempts that were
carried out without recourse to the central government. Fakhr al-Dīn
ʿUthmān, majordomo to the sultan al-Malik al-Kāmil, received the entire
Fayyum as an iqṭāʿ in 620/1223-4 CE. In the early decades of the seventh/
thirteenth century, the Fayyum was usually granted in its entirety as an
iqṭāʿ to members of the ruling family, contrary to the generally increasing
fragmentation of iqṭāʿs under the Ayyubids. Fakhr al-Dīn was granted
unconditionally (darbastā) the grain revenues (ḥawāṣil ), sugarcane, and
cattle of the Fayyum. In return, Fakhr al-Dīn committed to providing 200
horsemen, unspecied cash payments, and grain to the royal granaries. e
Ayyubid chronicler Makīn b. ʿAmīd, to whom we owe this information,
also extols the generosity of Fakhr al-Dīn, who is said to have built madra-
sas and mosques, as well as schools and endowments for orphans.33
Al-Nābulusī tells us that, once Fakhr al-Dīn became aware of the water
shortage in the province that had been granted to him, “he wished to dis-
play to him [the sultan] evidence of his eorts in all that he was in charge of,
and he looked for ways to bring prosperity to the Fayyum.” He then embarked
on lengthy and expensive attempts to repair the water supply of the Fayyum,
for which he received no support from the central government. His rst
attempt to deal with the water shortage was to clean up the Baḥr Yūsuf. At
his orders the banks of the canal were cleared of reeds, shrubs, and trees,
voir, see also G. Garbrecht “Historical Water Storage for Irrigation in the Fayyum Depres-
sion (Egypt),” Irrigation and Drainage Systems 10 (1996): 47-76.
31) For the water sources and tax revenues of these villages, see “e Database: Village
Details,” on the “Rural Society in Medieval Islam” website (Rapoport and Shahar).
32) is corrects Garbrecht’s conclusion that an important part of the dam (which survives
today) was not built until after 1245 (Garbercht, “Historical”: 65).
33) Makīn b. al-ʿAmīd (d. 672/1273), Taʾrīkh, ed. Cl. Cahen, Bulletin d’Études Orientales,
15 (1955): 133-4; see also Sato, State: 181.
12 Y. Rapoport, I. Shahar / JESHO 55 (2012) 1-31
with the objective, it seems, of preventing loss of water to the vegetation.
is, according to al-Nābulusī, “had no eect whatsoever, except for having
no trees [on the banks], and the greenery of the canal went away along with
the trees on its banks.”34
Fakhr al-Dīn’s second attempt to x the problem was a much more
ambitious and expensive project, an alteration to the dam at al-Lāhūn. As
noted above, silting in front of the embankment made the canal shallower
and caused more water to escape through the dam and back to the Nile.
Fakhr al-Dīn therefore decided to raise the height of the entire length of
the dam by a cubit and a half (1 metre), with the aim of diverting more
water to the Fayyum. We are not told how this major construction work
was executed, but it seems to have been nanced out of Fakhr al-Dīn’s cof-
fers; no corvée labour is mentioned.
Raising of the dam also had the unintended eect of increasing sedi-
mentation in front of the dam and at the entrance to the Fayyum itself.
Al-Nābulusī claims that the original height of the dam, as planned by
“ancient” pre-Islamic engineers, was precisely calculated to prevent the
problem of silting, but the engineers of al-Nābulusī’s time were not as
knowledgeable and disturbed the delicate balance. e continuous accu-
mulation of silt kept blocking the mouth of the Grand Canal entering the
Fayyum, the vital canal that owed through the Lāhūn gap. e only way
to allow enough water to ow into the Fayyum was for peasants from all
the villages of the Fayyum to gather before the inundation period began,
for removing huge amounts of silt:
e soil and the sand were held back until they formed mounds within the al-Munhā
canal [Baḥr Yūsuf ].35 And in the place from which the water enters to the Fayyum, in
front of its opening, a huge shoal (dikka) of soil was formed. is shoal becomes visible
every year in the Coptic month of Bashans (or Pashons, May 9-June 7), and the water
gets absorbed in it. So, men from all the villages of the Fayyum gather there and they cut
it with shovels, carrying it o with large baskets. en, the water bypasses the shoal from
its two sides, entering the canal that connects to the Fayyum and its villages through two
narrow mouths, one seven cubits wide and the other ve cubits wide. Both have a depth
of no more than two cubits.36
34) Al-Nābulusī, Taʾrīkh: 16.
35) Al-Nābulusī often refers to Baḥr Yūsuf as Baḥr al-Munhā.
36) Al-Nābulusī, Taʾrīkh: 16.
Irrigation in the Medieval Islamic Fayyum 13
Remarkable, again, is the lack of any ocial directive by the state to organ-
ize this annual cleaning of the mouth of the Grand Canal. e peasants of
the Fayyum apparently did this work voluntarily, in order to secure a suf-
cient supply of water for their elds.
Al-Nābulusī’s account of the raising of the dam and its unexpected conse-
quences is one of rudimentary engineering, which did not reach the standards
of the pre-Islamic sages. is theme is even more pronounced when he
describes Fakhr al-Dīn’s most ambitious attempts to improve water supply,
which involved cutting a new opening on the western bank of the Nile, at
the head of Baḥr Yūsuf, and erecting articial islands in the Nile itself.
First, at the advice of local experts, he travelled with tools and beasts of
burden to the mouth of the Baḥr Yūsuf, some 300 kilometres south of the
Fayyum, where he ordered a new opening to be cut about 350 metres north
of—i.e., downstream from—the old opening. But, again, this eort had
unexpected eects, with some of the water that used to enter through the old
opening escaping through the new opening back into the Nile, thus actually
reducing the ow in the Baḥr Yūsuf. Eventually, the new opening was blocked
with sand and clay, in two years.
Following this failure to increase the volume of water in Baḥr Yūsuf
through a new opening, the relentless Fakhr al-Dīn tried sinking several large
boats in the Nile itself, right o the head of the canal. e objective was to
form an articial island that would divert water into the opening of Baḥr
Yūsuf, but, because the boats were positioned incorrectly, the water was
diverted around it, and a huge island was formed at the head of the canal,
with the water of the Nile owing behind it and away from Baḥr Yūsuf.
According to al-Nābulusī, this spectacular failure caused a further decrease in
the amount of water available to the Fayyum, and it is only after the creation
of this articial island that the head of the Baḥr Yūsuf began to dry up for
eight months of every year.
e striking feature of this scathing and almost comic account of the
attempts to increase the water supply to the Fayyum is the decisive role of the
local villagers and their rustic knowledge. Al-Nābulusī says that Fakhr al-Dīn’s
advisers were all men from the village of al-Lāhūn who had no formal train-
ing but had evidently become local experts because of their knowledge of the
operation of the dam located next to their village. ey are the “so-called
engineers” (al-musammūn bi-ʾl-muhandisīn):
A group of people known as engineers, from among the people of the village called
al-Lāhūn. It [the village] lies in the vicinity of the aforementioned dam, and it is their
14 Y. Rapoport, I. Shahar / JESHO 55 (2012) 1-31
custom to oversee the operation of installing the “piece” that was mentioned above. For
that reason they are called engineers, not for their knowledge of engineering or for their
experience of it.37
ese villagers from al-Lāhūn are the ones who went with Fakhr al-Dīn to the
head of the Baḥr Yūsuf and advised him rst to divert to Nile water to a new
opening, then to sink boats in the Nile in order to create an articial island.
Fakhr al-Dīn did not bring with him professional engineers from Cairo. In
fact, the account gives a sense of collaboration between the iqṭāʿ holder and
the local peasantry in order to increase the prosperity of the province.
Why did these attempts fail? Overall, the picture that emerges is one of
ambitious and expensive attempts to solve the problem of water shortage in
the Fayyum. While al-Nābulusī blames the problem on negligence and lack
of investment and states that the central government had not spent any
money on maintaining the water supply of the Fayyum for over a century,
Fakhr al-Dīn ʿUthmān’s persistence demonstrates that funds for maintaining
the local irrigation system actually came from the local iqṭāʿ holder and not
from Cairo. If the Fayyum suered a water shortage, it was not due simply
to a lack of investment or commitment by the military elite.
Nonetheless, al-Nābulusī’s account suggests that the challenges facing
Fakhr al-Dīn were related to the decentralized and localized nature of the
irrigation system that developed in the Islamic period. First, Fakhr al-Dīn
lacked cadres of professional engineers trained in the capital, who would
be able to advise him technically on the grand projects he was undertaking.
He relied on the local knowledge of the peasantry, which was appropriate
to the upkeep of the irrigation system but not to its development. Second,
as noted above, al-Nābulusī believed that part of the problem of water sup-
ply in the Fayyum was caused by water of the Baḥr Yūsuf being drawn o
by upstream villages, in the provinces of al-Ashmūnayn and al-Bahnasā. A
solution to this conict of interests would have required coordinated plan-
ning from the centre, something that Fakhr al-Dīn, as a provincial iqṭāʿ
holder, could not have achieved on his own.
Water Allocation
e other key element of all irrigation systems, besides the water supply—
which was more important in the Fayyum than in other Egyptian prov-
inces—was the management and mechanisms of water allocation to
37) Al-Nābulusī, Taʾrīkh: 16.
Irrigation in the Medieval Islamic Fayyum 15
villages. Unlike the rest of the Egyptian countryside, the Fayyum depres-
sion was irrigated by gravity-fed canals, which enabled the spread of sum-
mer crops, plantations, and orchards. But gravity-fed canals, like small
rivers, cause friction between upstream and downstream communities.
Analysis of al-Nābulusī’s survey, however, shows that the water allocation
in the Fayyum was devised in a way that ensured downstream communi-
ties were compensated for the loss of water through seepage and evapora-
tion, known as conveyance loss. Most importantly, this system appears to
have been largely self-regulated, with minimal interference from the cen-
tral or provincial governments.
While almost all the villages in the Fayyum depression were irrigated by
gravity-fed canals, a minority of villages in the eastern part of the province
were watered only by the annual Nile ood, in the manner common to
the rest of Egypt. Al-Nābulusī describes these villages as being irrigated in
“the manner of the Rīf,” that is, Lower Egypt. As shown in Map 1, annual-
ood irrigation was used in villages that lay on the banks north and south
of the al-Lāhūn dam, as well as in villages along the arm of the Baḥr Yūsuf
that led water from the al-Lāhūn dam to the centre of the depression. It is
also mentioned as an alternative or additional method of irrigation in two
villages that were located on the eastern slopes of the depression.38 e vil-
lages irrigated in the “manner of the Rīf,” by the annual Nile oods that
take place in the late summer and early autumn, were necessarily limited
to winter crops, as al-Nābulusī repeatedly comments. Most of these were
small or medium-sized villages, specializing in the cultivation of ax; most
of the ax in the province was cultivated there.
As noted above, the vast majority of the villages in the Fayyum were
irrigated by gravity-fed canals that branched o the Grand Canal (al-khalīj
al-aʿẓam), the name used at the time for the arm of the Baḥr Yūsuf that
entered the depression. After the water was diverted by the Lāhūn dam
towards the Fayyum, the water entered the Grand Canal and owed from
southeast to northwest, towards Madinat al-Fayyum. A dense network of
38) ese are Bandīq and Dumūh al-Dāthir. See also the insightful comments on these vil-
lages by Ball, Contributions: 220. Bandīq is said to have been irrigated in the manner of the
Rīf but is also mentioned as being fed by the Wardān canal. is probably means that the
Wardān canal was used as a drainage canal in this period; a similar link between drainage
and irrigation by annual inundation is mentioned for the village of Dumūshiya, close to
Madinat al-Fayyum. Dumūh al-Dāthir is mentioned as a small village, where some land is
cultivated by the water of the Nile, as in Lower Egypt (al-Rīf ) and some of it by means of
waterwheels (saqy), as in the villages of the Fayyum.
16 Y. Rapoport, I. Shahar / JESHO 55 (2012) 1-31
Irrigation in the Medieval Islamic Fayyum 17
feeder canals branching out from the Grand Canal provided water for about
120 villages and hamlets in the depression.
e network of the main irrigation canals is recorded carefully in the
fth/eleventh-century account of Abū Iṣḥāq, who was interested mainly in
the infrastructure of the irrigation system rather than in the village com-
munities and their tax revenues. According to his account, most major
branch canals were controlled by systems of ancient, pre-Islamic sluice
gates (bāb), “dating from the time of Joseph,” with a capacity, i.e., width,
of two to three cubits. e Tanabṭawiya canal, branching o south of the
Fayyum, was said to be controlled by three sluice gates, each with a capac-
ity of two cubits. e Dilya (or Delahe) canal was said to be controlled by
two sluice gates, with a capacity of 2.25 cubits each. e next three major
canals, called Tlālh, Bamūh (Bamūya or Bimwa) and Tndh, were also con-
trolled by symmetrical pairs of sluice gates of 2.33, 2.5, and 2.25 cubits.39
Abū Iṣḥāq notes that almost all the canals that had sluice gates were
subject to a strict schedule of opening and closure and were therefore
known as al-muṭāṭiyya (literally, “alternating”). According to this schedule,
all the alternating canals were simultaneously closed and opened for peri-
ods of twenty days between November and April, before the driest season.40
e term is already mentioned in the third/ninth-century work of Ibn
ʿAbd al-Ḥakam. He reports that Joseph set up the irrigation of the Fayyum
so that “that ‘alternating’ [canals] will be diverted to upper-level [canals],
and vice versa, according to an hourly schedule, by night and day”
(wa-uṣayyiru muṭāṭiʾān li-l-murtaʿ wa-murtaʿān li-l-muṭāṭiʾ bi-awqāt
min al-sāʿāt fī l-layl wa-l-nahār).41 In his fourth/tenth-century account of
the Fayyum, al-Masʿūdī also mentions that the canals of the Fayyum fell
into several categories: upper-level (murtaʿ ), the alternating (muṭāṭīʾ) and
39) Al-Maqrīzī, al-Mawāʿiẓ, 1:671-4.
40) e schedule of these canals was as follows: they were closed from 10 Hatūr (20 Novem-
ber) until the end of that month (9 December); then open from the beginning of Kihāk
(10 December) for 20 days; then closed for the 10 remaining days of Kihāk, until Epiphany
(January 6), and opened on Epiphany, until the end of Ṭūbah (7 February); then they
were closed from the beginning of Amshir (8 February) for 20 days; then they were opened
for the 10 remaining days of Amshir (28 February), until the 20 Baramhat (29 March);
then they were closed for 30 days, until they were opened 10 days before the end of
Barmūda (28 April). Presumably the canals were then left open until the cycle began again in
November.
41) Ibn ʿAbd al-Ḥakam, Futūḥ Miṣr wa-l-Maghrib (Cairo: Maktabat al-aqāfa al-Dīniyya,
1995):14; repeated in al-Maqrīzī, al-Mawāʿiẓ, 1:665.
18 Y. Rapoport, I. Shahar / JESHO 55 (2012) 1-31
muṭāṭīʾ al-muṭāṭīʾ, which is, explains al-Masʿūdī, “a term used by the Egyp-
tians, meaning the lower (al-munkhaḍ ).”42 ese references from the
earliest Islamic geographical literature indicate that the annual schedule for
opening and closure of the “alternating” low-lying canals was intended to
prevent excessive ow of water during the winter months, in order to divert
some of it to canals crossing higher grounds.43
Al-Nābulusī’s survey is focused on the end users, the communities of
cultivators and their water rights, and he never mentions the alternating
schedule of canals. is is probably not the result of any dramatic change
in water allocation and management. e pre-Islamic methods of ensuring
that higher-level canals received sucient irrigation water would have been
as necessary in the seventh/thirteenth century as they were in the fth/
eleventh, but, because al-Nābulusī is interested mainly in the villages as
units of production, he says little about the management of the irrigation
system as a whole. Instead, his focus is on identifying the feeder canal that
irrigates each peasant community, and the water rights (ḥuqūq) of the
community or individuals within it.
e distinctive element of the irrigation system described by al-Nābulusī
was that most villages irrigated by gravity-fed canals were accorded a water
right, determined by the width of the weir at the head of the local feeder
canal.44 is water right was measured in st-lengths (qabḍas), each st-
length measuring one sixth of a cubit. Often, the allocation of water was
through dividers (maqsam, pl. maqāsim), which consisted of clusters of
weirs. ese dividers were used when more than one canal was splitting o
at the same spot, and they were found only on the Grand Canal and other
major canals (the baḥrs). In some cases, internal divisions of water quotas
within a village to certain crops or individual land owners were also speci-
ed: In the village of Dhāt al-Ṣafā, for example, ve qabḍas, out of the
31 qabḍas allocated to it, were designated for newly planted sugarcane.
ere are several more such examples in al-Nābulusī’s work, mostly with
the water being allocated to specied orchards or to the plots of village
headmen.45
42) Al-Masʿūdī, Murūj, 2:72 [no. 784].
43) On the relatively little we know about this system in the pre-Islamic period, see Ball
Contributions: 216.
44) For more technical details about the Fayyum weirs, see D. B. Kraatz and K. Mahjan,
Small Hydraulic Structures. FAO Irrigation and Drainage Paper 26/1. (Rome: Food and
Agriculture Organization of the United Nations, 1975).
45) See also the examples from al-Nābulusī’s survey, discussed in Sato, State: 222-4.
Irrigation in the Medieval Islamic Fayyum 19
is allocation of water rights by specifying the width of the weir at the
head of the local branch canal is mentioned in the third/ninth century by
Ibn ʿAbd al-Ḥakam, alongside muṭāṭiʾ canals. Ibn ʿAbd al-Ḥakam reports
that, after setting the schedule for opening and closing muṭāṭiʾ and murtaʿ
canals, Joseph also decreed that “each of the villages would be apportioned
st-lengths (qabaḍāt), so that none will receive less than it has a right to
(ḥaqqihi), nor will it receive more than what it is able (to use).”46 e terms
used in this early Islamic account are precisely the terms used by al-Nābulusī
in the seventh/thirteenth century.
Allocation of water rights was necessary only where access to gravity-fed
canals needed to be regulated. ose villages mentioned earlier, which were
irrigated solely by the annual ood, “in the manner of the Rīf,” were not
part of this network of water allocation. In addition, there was a minority
of villages irrigated by gravity-fed canals that had no specied water rights.
is meant that the local canal that irrigated the village had no weir at all,
or that the head of the weir had no specied width. is was an indication
of insucient water ow, which was in no need of regulation. In six of
these villages, al-Nābulusī uses the formula “canal without quota, due to
the elevation of the land” (bi-ghayr ʿibra bi-ḥukm ʿuluwwi al-arḍ ), meaning
that there was no need to regulate the ow of water into these canals, as it was
limited by topography.47
e actual water right of a village—measured by the width, in qabḍas,
of the opening at the head of the local feeder canal—was dependent on the
location of the village and its size. Map 2 clearly shows that the water
quota of a village was often related to the distance of the village from the
Grand Canal; the further downstream the village was along a branch canal,
the larger was its water quota. e pattern is very clear in the long Minyat
Aqnā canal, which fed the villages at the western end of the Fayyum. e
villages closest to the centre of the province were the village known as the
Dinfāras of Jaradū and Ihrīt and the village of Babīj Anqāsh. ese two
villages had quotas of four and 4.5 qabḍas respectively; the next village
downstream, al-Ḥanbūshiyya, had an opening of foourteen qabḍas; and
the large village of Minyat Aqnā, right at the western end of the cultivated
area, received its water from a feeder canal with an opening of fty qabḍas.
A similar pattern is seen in the Dhāt al-Ṣafāʾ canal, feeding villages in the
46) Al-Maqrīzī, al-Mawāʿiẓ, 1:665.
47) ese are the villages of Ṣanūfar, Ghābat Bāja, Qushūsh, Minyat Karbīs, Minyat al-Dīk,
Banū Majnūn, and Shalmaṣ (al-Nābulusī, Taʾrīkh: 126, 132, 143, 146, 165).
20 Y. Rapoport, I. Shahar / JESHO 55 (2012) 1-31
Irrigation in the Medieval Islamic Fayyum 21
northeastern Fayyum, where the rst village upstream had a right to an
opening of six qabḍas, while the sixth village downstream had a right to
31 qabḍas. e pattern is less obvious along the two large branch canals—
the Dilya and the Tanabt ̣awiya—that watered the southern Fayyum, where
some villages close to the centre had relatively large openings in their feeder
canals.48
e system of allocation in the Fayyum is of a type well known in other
Islamic societies, which Glick terms “Syrian.”49 In this system, the total
discharge of the river or spring was divided among the principal canals
taking water from it, in proportion to the amount of land served by each
canal. e water was assigned to the lands it irrigated and could not be
alienated or sold. e fth/eleventh-century jurist al-Māwardī refers to
this model as one of several methods for dividing water rights in rivers or
canals. He said that in this method the cultivators “divide the mouth of the
river with a board (khashaba) that stretches across both banks of the river.
e board is divided by gaps (ḥufūr) according to their water rights. In
each gap water enters according to what is due its owner, whether a fth or
a tenth.”50
While the method of water allocation is familiar, the actual distribution
of water rights in the Fayyum shows uniquely that downstream communi-
ties had rights to disproportionately wide openings in their feeder canals.
e reason for this was probably that the inevitability of conveyance
loss through evaporation and seepage.51 As the above discussion makes
clear, the width of the opening was not the only factor determining the
amount of water available to each of the villages. Some villages were never
allocated a set width to their feeder canal, because their lands were located
on higher ground, and there was no need to restrict them. Individual top-
ographical conditions must have been a factor in determining the width of
the local weir in other villages as well. But, overall, the system of rights was
intended to ensure that each of the villages along the gravity-fed canals
received a fair share of water and that downstream communities were not
disadvantaged.
48) For full details of the water rights of each of the villages, see “e Database: Village
Details,” in “Rural Society in Medieval Islam” website (Rapoport and Shahar).
49) Glick, Fortress: 79; Price, “Evolution”: 73.
50) Al-Māwardī, Kitāb al-Aḥkām al-sulṭāniyya wa-l-wilāyat al-dīniyya, ed. Samīr Muṣṭafā
Rabāb (Beirut: al-Maktaba al-ʿAṣriyya, 2000): 202-3.
51) On conveyance loss, see Price, “Evolution,” 323-56, and the sources cited there.
22 Y. Rapoport, I. Shahar / JESHO 55 (2012) 1-31
Canal Management and Maintenance
e mediaeval accounts of the irrigation system of the Fayyum indicate
two mechanisms that aimed at egalitarian distribution of water: the “alter-
nating” regime of low-level canals, diverting water to canals which lay
on higher grounds; and the allocation of “rights,” which ensured that
downstream communities were compensated for conveyance loss. Both of
these egalitarian mechanisms were known in early Islamic times, and were
probably pre-Islamic. While the preceding section explains the general
principles of water allocation, we have yet to identify the institutions
that governed the actual functioning of the system, year by year, season by
season. Who decided how many qabḍas a village would be entitled to?
Who coordinated the opening and closure of the low-lying alternating
canals? What were the mechanisms for resolution of conicts between
downstream and upstream communities, which must have occurred dur-
ing water shortages?
For historians of Greco-Roman, pre-Islamic Fayyum, the answer would
be the state or the provincial governor. In the papyri we nd a wide range
of irrigation ocials, corvée labour, a separate workforce of irrigators dis-
tinct from ordinary peasants, and strong central control.52 In the Fayyum
visited by al-Nābulusī in the seventh/thirteenth century, however, the
intervention of the state appears to have been minimal. ere are only
minimal fees associated with local irrigation maintenance, and they sup-
port simple irrigation administration and dredging tools, not a distinct
class of irrigation labourers. ere is no reference to corvée labour. Rather,
the management of water rights in mediaeval Islamic Fayyum was subject
to the authority of local iqṭāʿ holders, as well as tribes or clans, which nego-
tiated water rights and coordinated labour contributions.
52) On the nomarchs, or provincial governors, of the land-reclamation era in the early
Ptolemaic period, see ompson, “Irrigation”: 107-22; on the hydroparochia (lit., “water
supply”), a term that appears beginning in the fourth century CE, and professional irriga-
tors, see D. Bonneau, Le régime administratif de l’eau du Nil dans l’Egypte grecque, romaine
et byzantine (Leiden: Brill, 1993): 216; D. Rathbone, Economic Rationalism and Rural Soci-
ety in ird-Century A.D. Egypt: e Heroninos Archive and the Appianus Estate (Cambridge:
Cambridge University Press, 1991): 166. On corvée labour in Roman Fayyum, set at ve
days a year, see Price, “Evolution,” 176-8, and the sources cited there. Bonneau does, how-
ever, concede that the actual distribution of water in the Roman period was probably rele-
gated to customary laws (D. Bonneau, “Loi et coutume en Egypte: Un exemple, les marais
du Fayoum appelés drymoi,” JESHO 26/1 [1983]: 1-13, here 3-4).
Irrigation in the Medieval Islamic Fayyum 23
e mediaeval authors do not say explicitly who had the ultimate polit-
ical power to allocate water rights. Abū Iṣḥāq, who is not directly inter-
ested in the water allocation but only in the infrastructure, refers vaguely
to local, customary laws: “ey [the villages of the Fayyum] have a division
(qism), by which every locality gets its water share (shurb) justly, according
to their customary rules (al-qawānīn al-mashhūra ʿindahum).”53 Al-Nābulusī,
who does record the water rights of gravity-fed villages, does not explain
the mechanism for the allocation of these rights. He does relate, however,
one case of an iqṭāʿ holder who decided to deny water rights to a village. In
his account of the village of Bushṭā, along the Dilya canal in the southern
Fayyum, he reports:
is village was formerly large and populous, but when its people became impudent
(taʿaddā) and, in their numbers, prevailed over the iqṭāʿ holder, he reduced their right to
water and transferred it to other villages, whose people are obedient. It has now turned
into a small village, with few tenants, and spacious land that lies fallow every year, because
of the reallocation of the water from it to other villages.54
is clear case of an iqṭāʿ holder diverting water from one village to another
has been seized upon as a testimony of the absolute power of the iqṭāʿ
holder over the peasant communities in the mediaeval Fayyum and the
Middle East in general.55 For our purpose, the case does show that iqṭāʿ
holders were able to intervene in the allocation of water rights, with serious
consequences for the local community, but the implications of this exam-
ple are limited by its circumstances. e village of Bushṭā formed part of a
much larger iqṭāʿ scal unit, the iqṭāʿ of the Dilya canal, which included at
least ten other villages. is was the largest iqṭāʿ unit in the province, and
it stands out against the backdrop of the vast majority of iqṭāʿ units, com-
posed of individual villages. is is crucial, because an iqṭāʿ holder who
deprived his village of water was also depriving himself of scal revenue.
e iqṭāʿ holder of the Dilya canal could aord to do that, because the
water diverted from Bushṭā went on to irrigate other villages within the
large iqṭāʿ unit. But ocers who held an iqṭāʿ of just one village were not
in a position to intervene in the allocation of water rights to the village—
unable to reduce it, because they would lose revenue, and unable to increase
53) Al-Maqrīzī, al-Mawāʿiẓ, 1:674.
54) Al-Nābulusī, Taʾrīkh: 65-6.
55) Sato, State: 233.
24 Y. Rapoport, I. Shahar / JESHO 55 (2012) 1-31
it, for that would divert water from villages that belong to the iqṭāʿ units of
other ocers.
By and large, the iqṭāʿ system created mutual dependency between iqṭāʿ
holders and peasant communities with regard to irrigation works, and it
encouraged a great deal of decentralization. For Egypt as a whole, Ibn
Mammātī explains that the maintenance of irrigation was divided into
major dykes and local dykes. e major dykes were the responsibility of
the sultan but were funded by fees (rusūm) from the provinces, collected
by agents of the central government. e local dykes, says Ibn Mammātī,
were the joint responsibility of local peasants and iqṭāʿ holders, who spent
their own funds on it.56 is general description applies to lands irrigated
directly by Nile oods, but the principle of collaboration between iqṭāʿ hold-
ers and peasants on the local level would have applied in the Fayyum too.
Had the irrigation system been controlled closely by the state, we would
expect to see indications of this in the tax obligations of the villages. In
fact, taxes specically designated for irrigation purposes indicate a low level
of state intervention in the management of irrigation. ere are three such
taxes recorded by al-Nābulusī. First, about half the villages along gravity-
fed canals—but not villages fed directly by the Nile oods—made small
payments, in kind, to the ocial known as khawlī al-baḥr (canal control-
ler). ese annual payments ranged from one-half to three ardebs (an ardeb
is about ninety litres) of grain, usually wheat. e total for 47 villages was
just under sixty ardebs for an entire year.57 ese fees were clearly intended
to support this individual ocial and his household and could not have
supported a signicant contingent of irrigation workers.
A second irrigation tax was known as the “raking” or “dredging” fee
(rasm al-jarārīf ). e jarrāfa (pl. jarārīf ), was a simple tool, measuring
roughly a metre on a side, used for the annual dredging of canals in Mam-
luk and Ottoman Egypt.58 Like the payments to the khawlī al-baḥr, the
dredging fee was paid by villages along gravity-fed canals but not by vil-
lages fed directly by the Nile ood. e total paid for 62 villages was just
under 4700 silver dirhems. e payment per village seems closely related
to the size of the village.59 Like the payments to the khawlī al-baḥr, the
56) Ibn Mammātī, Qawānīn, 232-3.
57) For full details of the payments to the khawlī al-baḥr, see “e Database: Miscellaneous
Taxes,” on the “Rural Society in Medieval Islam” website (Rapoport and Shahar).
58) Borsch, Black Death, 34-9.
59) e 24 gravity-fed villages that al-Nābulusī described as “large” paid 2303.5 dirhems,
just under 100 dirhems per village; the 22 “medium-size” villages paid 1069.25, about
Irrigation in the Medieval Islamic Fayyum 25
dredging fee was directly related to the management of the gravity-fed
canals and was paid locally, not to the iqṭāʿ holder. is tax paid for the
construction and upkeep of the jarārīf. ese taxes likewise could not have
sustained a workforce.
A third irrigation-related tax, mentioned at the end of al-Nābulusī’s sur-
vey, is a levy of just over a hundred dredging tools (jarārīf ) for the con-
struction of a dike in the district of Giza, outside of the Fayyum.60
Al-Nābulusī lists the contribution of each village in the Fayyum, measured
in the number of units (qiṭaʿ) of dredging tools, with many villages obliged
to provide fractions. Al-Nābulusī reports that out of 104 jarrāfa units listed
in the decree, the villages actually provided 95. Almost all villages in the
Fayyum were expected to participate, and there is no distinction between
gravity-fed villages and those irrigated by the Nile ood. e number of
units per village is, as with the local dredging fee, closely related to the size
of the village.61 e interesting feature of this levy is that the sultan expected
the local communities of the Fayyum to provide the dredging tools for the
work in Giza; he did not, as far as we can tell, demand corvée labour but
rather a contribution in tools, which were produced locally.
In the near absence of state intervention, the institution that most likely
took upon itself the management and the upkeep of the canal system was
the tribal group. While the population of the villages and towns of the
Fayyum consisted almost entirely of settled peasants, these settled com-
munities were divided into sedentary inhabitants (ḥaḍar) and Arabs
(ʿarab). e number of communities described as sedentary in al-Nābulusī’s
Fayyum was small, and they lived in only a handful of villages. e Arab
tribes constituted the vast majority of the population, inhabiting more
than ninety villages in the Fayyum. e Arabs were divided into three large
tribal confederacies: the Banū Kilāb, the Banū ʿAjlān, and the Lawāta. e
Banū Kilāb resided in the western part of the depression, the Banū ʿAjlān
in the east, and the Lawāta at the Lāhūn gap and at the entrance to the
depression. Each of these tribes was subdivided into several clans (see Map 3).
50 dirhems per village; and the 29 “small” villages paid 599.375 dirhems, an average of
20 dirhems. For full details of the fees, see “e Database: Miscellaneous Taxes,” on the
“Rural Society in Medieval Islam” website (Rapoport and Shahar).
60) Al-Nābulusī, Ta ʾrīkh: 178-9.
61) e 25 “large” villages were expected to provide 57.3 units, the 25 “medium-size” vil-
lages 20 units, and the 35 “small” villages 11.2 units. is does not add up to 104 units,
because the size of some villages is not specied by al-Nābulusī. For full details of the fees,
see “e Database: Miscellaneous Taxes,” on the “Rural Society in Medieval Islam” website
(Rapoport and Shahar).
26 Y. Rapoport, I. Shahar / JESHO 55 (2012) 1-31
A comparison between the distribution of the tribal groups and the
irrigation system of the Fayyum (Map 1 above) suggests a close correlation
between certain tribal groups and irrigation districts. In the case of some
particular branch canals, there is a denite overlap with a local clan or
subdivision of a tribe. All ve villages lying along the al-Sharqiyya canal, in
the water-scarce area of the eastern Fayyum, were inhabited by the Banū
Zarʿa clan of the ʿAjlān tribe. e Banū Zarʿa also occupied ten other vil-
lages in the eastern Fayyum, and their predominance in the region would
have allowed water management across several smaller canals. Eight of ten
villages along the Sinnūris canal were inhabited by the Qayāṣira, or the
Banū Qayṣar, of the ʿAjlān tribe. No other villages in the Fayyum were
inhabited by the Qayāṣira, a case of nearly complete overlap between a
branch canal and a tribal group. e pattern is also marked in the southern
part of the province: of a total of sixteen villages along the Dilya canal,
eight were occupied by the Banū Ghuṣayn of the Kilāb tribe, and ve of
nine villages along the Tanabṭawiya canal were inhabited by the Banū
Ḥātim, who also belonged to the Kilāb.
As is evident from the comparison of the two maps, it was not only that
individual irrigation canals were dominated by specic clans: Practically all
villages that were irrigated by the Nile ood were inhabited by the clans,
or subdivisions, of the Lawāta tribe; the Lawāta, on the other hand, did
not inhabit any of the villages irrigated by gravity-fed canals. ese gravity-
fed villages, including villages in the eastern Fayyum that were seasonally
irrigated directly by Nile oods, were divided between the ʿAjlān and the
Kilāb tribes, but the ʿAjlān and the Kilāb shared no main branch-canals on
the Fayyum. us, all the villages along the al-Sharqiyya, Dhāt al-Ṣafāʾ,
and Sinnūris canals were inhabited by the ʿAjlān; all the villages along the
southern Tanabṭawiyya, Dilya, and Minyat Aqnā canals were inhabited by
the Kilāb.
e overlap between tribal groups and irrigation canals is surely not
coincidental. As we have seen, allocation and administration of water
rights in mediaeval al-Andalus was in accordance with tribal norms. is
is a feature of many irrigation systems, where a rigid social ordering of
space gives rise to self-sucient communities, displaying a political unity
and an egalitarian ethic, with mechanisms for retaining exclusive control
of water.62 Anthropological studies of peasant tribal communities in mod-
ern Iraq, Yemen, and Oman examine the ways in which tribal groups of
62) Glick, Fortress: 69-73.
Irrigation in the Medieval Islamic Fayyum 27
28 Y. Rapoport, I. Shahar / JESHO 55 (2012) 1-31
settled peasants try—not always successfully—to mobilize labour and
resolve conicts between downstream and upstream communities.63
In the absence of strong state institutions for the maintenance and
upkeep of irrigation canals, the tribal segments in the mediaeval Fayyum
would have oered a readily available alternative. At the level of the local
branch canal, the dominant clan would perhaps have oered a mechanism
for resolving disputes between upstream and downstream communities
along the gravity-fed canals of the Fayyum, as well as a way of retaining
within the group its rights to the canal waters. At the level of the province
as a whole, it seems likely that the tribal confederacies of the ʿAjlān and the
Kilāb oered an institution for managing the allocation of water between
the dierent branch canals, as between the low-lying muṭāṭiʾ and the
upper-level murtaʿ.
Conclusion
e irrigation system of the mediaeval Islamic Fayyum, as recorded by
al-Nābulusī, was locally run and tribally controlled, in a manner much dif-
ferent from the centralized models often suggested for pre-modern Egypt.
is was much to the chagrin of al-Nābulusī himself, who, as a state
bureaucrat dispatched from Cairo, believed that irrigation solutions should
come from the centre, and who considered the local Fayyumis primitive
rustics. Like many other mediaeval Muslim authors, al-Nābulusī is writing
within a well-established conceptual framework, which tends to contrast a
legendary glorious past with a perceived corrupt present.64 A closer reading
of his account reveals, in fact, a system that was at least reasonably main-
tained, and perhaps even developed, during the rst half of the seventh/
thirteenth century. We have noted the numerous attempts by Fakhr al-Dīn
ʿUthmān to improve the water supply. New villages were established next
to the ones abandoned or in the centre of the depression, on lands that
became available after the contraction of Lake Qārūn. ere are many
other indicators of maintenance and development of the irrigation system.
Newly planted orchards and sugar plantations are reported in the villages
63) Varisco, “Sayl and ghayl ”: 379.
64) For a discussion of this tendency, see N. Matar, “Confronting Decline in Early Modern
Arabic ought,” Journal of Early Modern History 9 (2005): 51-78.
Irrigation in the Medieval Islamic Fayyum 29
of Maṭar Ṭaris, Fanū, and Saynarū; a new canal provided water to Talīt, a
village that relocated to a new site; a new sluice gate was built on the Dhāt
al-Safāʾ canal; and a new sugar press was installed in Sanhūr and a new mill
in Abū Ksā, both of which required an ample water supply.
Locally controlled and managed, the irrigation system did function. e
upkeep and regulation of the pre-Islamic al-Lāhūn dam continued but was
now based on local knowledge and materials, apparently without the involve-
ment of state ocials or corvée labour. Mobilization of labour for mainte-
nance work on the dam appears to have been organized locally and
voluntarily. Funds for maintaining the local irrigation system came from the
local iqṭāʿ holder and not from Cairo. If the Fayyum suered a water short-
age, it was not due simply to lack of investment or to a lack of commit-
ment by the military elite. is localization of knowledge and control was,
however, not without its faults: it was probably related to the abandonment
of the more labour-intensive sluice gates and weirs on the al-Lāhūn dam.
e lack of central planning and coordination may well have contributed
to a growing water shortage. Yet, the irrigation system of the Fayyum did
not collapse—far from it.
In line with the localization of irrigation knowledge and control, the
allocation of water rights in mediaeval Fayyum was remarkably egalitarian.
e “alternating” regime of low-level canals and the allocation of water
rights, measured by the maximum width at the head of the village feeder-
canal, ensured that both higher-elevation villages and downstream com-
munities received a fair share of the water. Both of these mechanisms are
attested in the third/ninth-century work of Ibn ʿAbd al-Ḥakam and may
well be a pre-Islamic legacy. Al-Nābulusī’s survey also strongly suggests
that these mechanisms were managed locally, with little interference from
the state. A single local ocial, the controller of the canal (khawlī al-baḥr),
received small fees in kind from villages that lay along the gravity-fed
canals; the same villages also made small contributions to the construction
of dredging equipment. But control of access to water was, overall, invested
in the local iqṭāʿ holders, in cooperation with the omnipresent tribal groups
that inhabited villages along individual canals. In the absence of central-
ized control, it was the clan, and above it the tribal confederacy, that seem
to have negotiated water rights and coordinated labour contributions.
Despite the stark dierences in topography and history, al-Nābulusī’s
Fayyum was as tribal as the rural parts of al-Andalus, suggesting a far-
reaching pattern of Islamic history that has yet to be fully explored.
30 Y. Rapoport, I. Shahar / JESHO 55 (2012) 1-31
Bibliography
Bagnall, Roger S., and Dominic W. Rathbone. 2004. Egypt: From Alexander to the Copts: An
Archaeological and Historical Guide. London: British Museum Press.
Ball, John. 1939. Contributions to the Geography of Egypt. Cairo (Bulaq): Government Press,
Survey and Mines Dept.
Boak, A. E. R. 1926. Irrigation and Population in the Faiyum, the Garden of Egypt. e
Geographical Review 16/3: 353-64.
Bonneau, Danielle. 1983. Loi et coutume en Egypte: Un exemple, les Marais du Fayoum
appeles drymoi. JESHO 26: 3-4.
——. 1993. Le régime administratif de l’eau du Nil dans l’Égypte grecque, romaine et byzan-
tine. Leiden: Brill.
Borsch, Stuart. 2004. Environment and Population: e Collapse of Large Irrigation Sys-
tems Reconsidered. Comparative Studies in Society & History 46: 451-68.
——. 2005. e Black Death in Egypt and England. A Comparative Study. Austin: Univer-
sity of Texas Press.
Clarysse, Willy. 2011. e Fayum: A First Introduction, in e Fayum Project (http://www
.trismegistos.org/fayum/index.php) [accessed 22 July 2011].
Cooper, Richard, S. 1973. Ibn Mammati’s Rules for the Ministries. Ph.D. dissertation, Uni-
versity of California, Berkeley.
Christiansen, Peter. 1993. e Decline of Iranshahr: Irrigation and Environment in the His-
tory of the Middle East 500 B.C. to A.D. 1500. Copenhagen: Museum Tusculanum
Press.
Decker, Michael. 2009. Plants and Progress: Rethinking the Islamic Agricultural Revolu-
tion. Journal of World History 20/2: 187-206.
Garbrecht, Günther. 1996. Historical Water Storage for Irrigation in the Fayyum Depres-
sion (Egypt). Irrigation and Drainage Systems 10: 47-76.
Glick, omas F. 1992. Hydraulic technology in al-Andalus. In e Legacy of Muslim Spain,
ed. S. K. Jayyusi. Leiden: Brill: 974-86.
——. 1995. From Muslim Fortress to Christian Castle: Social and Cultural Change in Medi-
eval Spain. Manchester: Manchester University Press.
Hewison, R. Neil. e Fayoum. History and Guide. Cairo: e American University of
Cairo Press, 2001.
Ibn ʿAbd al-Ḥakam. 1995. Futūḥ Miṣr wa-l-Maghrib. Cairo: Maktabat al-aqāfa
al-Dīniyya.
Ibn Mammātī. 1943. Qawānīn al-dawāwīn. Cairo: al-Jamʿīyah al-Zirāʿiyya al-Malikiyya.
Keenan, James, G. 1999. Fayyum Agriculture at the End of the Ayyubid Era: Nabulsi’s
Survey. In Alan K. Bowman and Eugene Rogan (eds.), Agriculture in Egypt from Phara-
onic to Modern Times (Proceedings of the British Academy 96). Oxford: Oxford Univer-
sity Press: 287-99.
——. 2003. Deserted Villages: From the Ancient to the Medieval Fayyum. Bulletin of the
American Society of Papyrologists 40: 119-40.
_____. 2005. Landscape and Memory: Al-Nabulsi’s Taʾrikh al-Fayyum. Bulletin of the
American Society of Papyrologists 42: 203-12.
Kirby, C., and D. Rathbone. 1996. Kom Talit: e Rise and Fall of a Greek Town in the
Fayium. Egyptian Archeology 8: 29-31.
Irrigation in the Medieval Islamic Fayyum 31
Kraatz, D. B., and I. K. Mahjan. 1975. Small Hydraulic Structures. FAO Irrigation and
Drainage Paper 26/1. Rome: Food and Agriculture Organization of the United
Nations.
Kraemer, Bryan. 2010. e Meandering Identity of a Fayum Canal: e Henet of Moeris /
Dioryx Kleonos / Bahr Wardan / Abdul Wahbi. Proceedings of the Twenty-Fifth Inter-
national Congress of Papyrology, Ann Arbor 2007, American Studies in Papyrology:
365-76.
Makīn b. al-ʿAmīd. 1955. Taʾrīkh, ed. Cl. Cahen. Bulletin d’études orientales, 15: 121-40.
al-Maqrīzī, Aḥmad b. ʿAlī. 2002-4. al-Mawāʿiẓ wa-l-iʿtibār fī dhikr al-khiṭaṭ wa-l-āthār, ed.
A. F. Sayyid. London: Muʾassasat al-Furqān li-l-Turāth al-Islāmī.
al-Masʿūdī, ʿAlī b. Ḥusayn. 1965-79. Murūj al-dhahab wa-maʿādin al-jawhar, ed. Ch.
Pellat. 7 vols. Beirut: al-Jāmiʿa al-Lubnāniyya.
Matar, Nabil. 2005. Confronting Decline in Early Modern Arabic ought. Journal of
Early Modern History, 9: 51-78.
al-Māwardī, ʿAlī b. Muḥammad. 2000. Kitāb al-Aḥkām al-sulṭāniyya wa-l-wilāyat al-dīniyya,
ed. Samīr Muṣṭafā Rabāb. Beirut: al-Maktaba al-ʿAṣriyya.
Mehringer, Peter J., Jr., Kenneth L. Peterson, and Fekri A. Hassan. 1979. A Pollen Record
from Birket Qarun and the Recent History of the Fayum. Quaternary Research 11:
238-56.
Mikhail, Alan. 2010. An Irrigated Empire: e View from Ottoman Fayyum. International
Journal of Middle East Studies 42: 569-90.
al-Muqaddasī, Shams al-Dīn Abū ʿAbdallāh. 1906. Kitāb aḥsan al-taqāsīm fī maʿrifat
al-āqālīm. Leiden: Brill.
al-Nābulusī, Abū ʿUthmān. 1898. Taʾrīkh al-Fayyum wa-bilādihi, ed. B. Moritz. Cairo:
Publications de la Bibliothèque Khédiviale.
Price, David H. 1993. e Evolution of Irrigation in Egypt’s Fayoum Oasis: State, Village and
Conveyance Loss. PhD diss., University of Florida.
Rapoport, Yossef, and Ido Shahar, Rural Society in Medieval Islam (http://www.history.
qmul.ac.uk/ruralsocietyislam) [accessed 22 July 2011].
Rathbone, Dominic. 1991. Economic Rationalism and Rural Society in ird-Century A.D.
Egypt: e Heroninos Archive and the Appianus Estate. Cambridge: e University of
Cambridge Press.
Sato, Tsugikata. 1997. State and Rural Society in Medieval Islam. Leiden: Brill.
Shafei, Ali. 1940. Fayoum Irrigation as Described by Nabulsi in 1245 A.D. Bulletin de la
Société Géographique Royal d’Égypte 20/3: 283-327.
——. 1960. Lake Moeris and Lahun Mi-wer and Ro-hun: e Great Nile Control Project
Executed by the Ancient Egyptians. Société d’Égypte Bulletin 33:187-215.
Varisco, Daniel, M. 1983. Sayl and Ghayl: e Ecology of Water Allocation in Yemen.
Human Ecology 11: 365-83 (reprinted in his Medieval folk astronomy and agriculture in
Arabia and the Yemen. London: Ashgate, 1997).
Wittfogel, Karl. 1957. Oriental Despotism: A Comparative Study of Total Power. New Haven:
Yale University Press.