Al-ʿUṣūr al-Wusṭā 30 (2022): 235-266

The Mosque‒Dār al-Imāra Complex at ʿAnjar:          
Preliminary Notes from a Multi-layered Exploration of 

Ceremonial Spaces in the Marwānid Period*

AilA SAnti

SOAS University of London

(as194@soas.ac.uk)

Abstract
Located in a corner of the Biqāʿ Valley, midway along the ancient route linking Beirut to Damascus, ʿAnjar 
remains one of the most significant—though least understood—Umayyad foundations in Greater Syria. With 
a plan shaped in strict accordance with the criteria of Hippodamian urbanism, the town constitutes a unique 
trait d’union between the classical urban tradition and the foundations of the early Islamic empire. This 
notwithstanding, and despite the importance of the site having been recognized and emphasized by many 
scholars, its origins, history, purpose, and patronage remain major enigmas in the field. This paper will discuss 
some insights surfaced from an analysis of the mosque‒dār al-imāra complex in ʿ Anjar with the aim to contribute 
to a more nuanced understanding of the site’s use and meaning in the Marwānid period. 

Introduction 

In the regional panorama of towns and extra-urban settlements dating to the first century 
of Islamic history, the site of ʿAnjar stands out for several unmatched characteristics.   
From its location (on the fringes of a large marshy lake),1 to its size2 (too big for a qaṣr 

* This article stems from the research I conducted as an Andrew Mellon Postdoctoral Fellow at the Center 
for Arts and Humanities, Faculty of Arts and Science, of the American University of Beirut in the Academic 
Year 2019–20. I would like to express my gratitude to Prof. John Meloy, Prof. Abdulrahim Abu Husayn, and Prof. 
Hermann Genz for their constant support and kind help during that challenging period. I wish also to thank 
Prof. Fabio Fabiani (Università di Pisa) for discussing with me the preliminary results of the survey and for the 
valuable insights he gave me.

1.  The existence of extensive wetland in close proximity to the site is documented as early as the fourth 
century BCE by the Greek historian Theophrastus, who described a large lake located in a mild depression 
between Mount Lebanon and another low mountain, where sweet flag grew in the summertime (Theophrastus, 
Enquiry into Plants and Minor Works, vol. 2, On Odours. Weather Signs, trans. A. Hort [London: W. Heinemann; 
New York: G. P. Putnam’s Sons], 247–8). A new predictive model of these wetlands—which lasted in the area 

© 2022 Aila Santi. This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License, 
which allows users to copy and distribute the material in any medium or format in unadapted form only, for noncommercial purposes only, and only so long 
as attribution is given to the original authors and source.

mailto:as194%40soas.ac.uk%20?subject=


236  •  AilA SAnti

Al-ʿUṣūr al-Wusṭā 30 (2022)

but too small for a proper madīna),2to its strongly classical-looking appearance, and the 
unbalanced ratio in its zoning pattern,3 everything seems to prevent ʿAnjar from fitting into 
any of the established categories of early Islamic urbanism. 

A few attempts have been made to explain the enigmatic nature of the site. Among them, 
that by Robert Hillenbrand4 remains the most comprehensive so far produced. After trying 
to frame ʿAnjar within virtually any possible settlement type from Roman antiquity to early 
Islam—suggesting it to have been a late Roman colonia, a very large country residence, an 
administrative center of the Umayyad peripatetic court, a place of refuge, a hunting lodge, 

an “instrument of segregation”5 intended to accommodate and separate a specific tribal 
grouping from the others, a garrison point, or a market-city—Hillenbrand concluded that 
none of these models alone can yield a full understanding of the site.6 

However, despite having exhausted the range of possible labels to attach to ʿAnjar, 
Hillenbrand has undoubtedly provided the most thorough possible interpretation of the 
site’s purpose and meaning, even if he has not entirely solved the enigma of its unique 
nature. What is made clear by his analysis of the site is that ʿAnjar must be viewed within the 
wider program of the systematic colonization of the Syrian countryside by the Umayyads,7 

until the intensive Mamluk drainage work of the eighth/fourteenth century—has been recently established by 
cross-referencing the geographical data provided by textual sources with remote sensing surveys and GIS. See 
G. Abou Diwan and J. Doumit, “Ancient Wetlands of the Biqāʿ: A Buffer Zone between the Hinterlands of Sidon 
and Berytus in the Roman Period,” Bulletin d’archéologie et d’architecture libanaises 16 (2016): 215–52.

2.  ʿAnjar’s enclosure measures 370 by 310 meters, making the site three times larger than the largest of 
the quṣūr, the so called “Large Enclosure” at Qaṣr al-Ḥaīr al-Sharqī, measuring 167 meters in length; see D. 
Genequand, “From ‘Desert Castle’ to Medieval Town: Qasr al-Hayr al-Sharqi (Syria),” Antiquity 79, no. 304 (2005): 
350–61, at 351.

3.  In particular, the fact that the housing was confined to only one of the four quadrants. Hillenbrand has 
calculated that the housing quadrant might have accommodated no more than twenty-six households; see R. 
Hillenbrand, “ʿAnjar and Early Islamic Urbanism,” in The Idea and Ideal of the Town between Late Antiquity 
and the Early Middle Ages, ed. G. P. Brogiolo and B. Ward-Perkins, 58–98 (Leiden: Brill, 1999), 75. Barbara Finster 
has remarked that without stratigraphic evidence—of which there is none—it is impossible to precisely date 
the houses built intra muros in ʿAnjar; see B. Finster, “ʿAnjar: spätantik,” in Residences, Castles, Settlements: 
Transformation Processes from Late Antiquity to Early Islam in Bilad al-Sham, Proceedings of the International 
Conference held at Damascus, 5-9 November 2006, ed. K. Bartl and A. R. Moaz, 229–42 (Rahden: Verlag Marie 
Leidorf, 2008), 229. In a recent article, Beatrice Leal has reported abundant evidence of the multiphasic nature 
of the so-called residential quadrant in ʿAnjar, attesting that some of the houses were turned into shops, 
workshops, and even a church at a later, conceivably post-Umayyad, time; see B. Leal, “ʿAnjar: An Umayyad 
Image of Urbanism and Its Afterlife,” in Encounters, Excavations and Argosies: Essays for Richard Hodges, ed. 
J. Moreland, J. Mitchell, and B. Leal, 172–89 (Oxford: Archaeopress, 2017), 182–4. These later superimpositions, 
however, do not disprove the fact that the original quadrant was planned and built as an integral part of 
the Umayyad project. In a recent monograph, Giuseppe Labisi has shown that the architectural and spatial 
characteristics of the residential insulae in ʿAnjar, along with the evidence provided by the graffiti found in 
situ, constitute, in fact, enough proof to consider them Umayyad in dating; see G. Labisi, Dwelling Models of 
Umayyad Madāʾin and Quṣūr in Greater Syria (Oxford: BAR Publishers, 2020), 24–31.

4.  Hillenbrand, “ʿAnjar.”
5.  Idem, “ʿAnjar,” 88.
6.  Idem, “ʿAnjar.”
7.  Idem, “ʿAnjar,” 91.



Al-ʿUṣūr al-Wusṭā 30 (2022)

The Mosque‒Dār al-Imāra Complex at ʿAnjar •  237

even though, as he warns, the current limited state of knowledge prevents the formulation 
of a more precise hypothesis in this regard.

More recently, Beatrice Leal has interpreted ʿAnjar as an imperial show-capital that 
also functioned as an important trading center.8 Expanding upon one of Hillenbrand’s 
suggestions, Leal argued that the plethora of shops at the site, far from being a mere 
“unthinking replication of a classical model,”9 were in fact used for commercial purposes 
and constituted a precise reference to the dynasty’s promotion and control of trade.10 In her 
view, the emphasis on the commercial dimension of ʿAnjar should be understood in terms 
of a demonstration of the wealth and economic strength of the empire: two crucial aspects 
in the construction of imperial identity in the Marwānid period. 11 

Although these studies have shed some light upon ʿAnjar’s meaning and purpose and 
revealed the multilayered complexity inherent to the site’s nature, they inevitably remain 
tentative and partial in the light of the fragmentary state of knowledge of the site. Indeed, 
even though the Umayyad ruins have been the subject of prolonged studies,12 the lack of 
a comprehensive publication of the results of their excavation,13 the highly speculative 
restoration works the site underwent, and the fact that its ceramic material was never 
studied and not even partially published (and that it mysteriously vanished along with 
the other archaeological finds during the Lebanese civil war) represent insurmountable 
obstacles to an adequate understanding of the site. Moreover, apart from the pioneering 
and very general work of Leon Marfoe in the 1970s,14 this part of the Biqāʿ Valley has never 
been the subject of a systematic archaeological survey,15 with the result that the territory 

8.  Leal, “ʿAnjar.”
9.  Hillenbrand, “ʿAnjar,” 97.
10.  Leal, “ʿAnjar,” 180.
11.  Idem, “ʿAnjar,” 181.
12.  The ruins were first documented by the Deutsche Baalbek Expedition (1901–4) and identified with the 

remains of Chalcis ad Libanum; see D. Krencker and W. Zschietzschmann, Römisch Tempel in Syrien (Berlin: 
de Gruyter, 1938), 182–91. For many years, scholars agreed on this identification (J. P. Rey-Coquais, “Notes de 
géographie syrienne antique,” Mélanges de l’Université Saint-Joseph 40 [1964]: 287–312, at 290; G. Schmitt, 
“Zum Königreich Chalkis,” Zeitschrift des Deutschen Palästina-Vereins 98 [1982]: 110–24, at 111; H. Seyrig, , “Les 
dieux syriens en habit militaire,” Annales archéologiques arabes syriennes 21 [1971]: 67–70) until Jean Sauvaget 
recognized the site as the Umayyad foundation of ʿAnjar on the basis of sound textual and archaeological 
evidence; see J. Sauvaget, “Les ruines omeyyades de ʿAndjar,” Bulletin du Musée de Beyrouth 3 (1939): 5–11. See 
also idem “Notes de topographie omeyyade,” Syria 24 (1944–5): 96–112. The site was thoroughly excavated—and 
almost simultaneously heavily rebuilt—between 1953 and 1975 by the Direction Générale des Antiquités du 
Liban under the guide of archaeologist Maurice Chehab and architect Haroutun Kalayan.

13.  No report was ever published of the over twenty years of archaeological investigations. Maurice Chehab 
only authored a very short article focused on the Great Palace (M. Chehab, “The Umayyad Palace at ʿAnjar,” 
Ars Orientalis 5 [1963], 17–25), while the first comprehensive discussion and complete plan of the remains was 
published in 1979 by Keppel Archibald Cameron Creswell (K.A.C. Creswell, Early Muslim Architecture: Umayyads, 
A.D. 622–750, 2nd ed., vol. 1, part 2 [New York: Hacker Art Books, 1979], 477–81).

14.  L. Marfoe, Between Qadesh and Kumidi: A History of Frontier Settlement and Land Use in the Biqa, 
Lebanon (PhD diss., University of Chicago, 1978).

15.  Recent survey work concentrated on the Baalbek Hinterland (B. Fischer-Genz and H. Ehrig, “First Results 
of the Archaeological Survey Project in the Territory of Ancient Heliopolis-Baalbek,” Bulletin d’archéologie 



238  •  AilA SAnti

Al-ʿUṣūr al-Wusṭā 30 (2022)

and context of ʿAnjar remain almost entirely unknown. This situation has forced scholars 
like Barbara Finster—who conducted research at the site in the early 2000s16 —to restrict 
the focus of their studies to the material aspect of the ruins which, abstracted from their 
stratigraphic context and heavily compromised by controversial modern rebuildings lend 
themselves to essentially descriptive research.

In order to address some issues stemming from the analysis of the published material, 
in 2017 I carried out a first short survey at the site. This preliminary fieldwork centered 
on the mosque site and highlighted a set of evidence—which I have discussed in a recent 
article17—that suggest that the building site, contrary to prevailing beliefs, underwent a set 
of complex architectural events. This, along with the remarkable abundance of late antique 
spolia, and hints found in the written sources,18 provided a basis for speculation that the 
Umayyad madīna might not have been founded entirely on virgin soil.

Following this lead, in 2020 I conducted further research in ʿAnjar as part of a postdoctoral 
residence at the American University of Beirut. Although not confirming a pre-Islamic date 
for the site, these fieldwork activities19 produced a fresh body of data that suggests that the 
Islamic settlement may not be the result of a unitary project. This paper will not present in 
detail the results of this second survey (which  I plan to publish in full in the future),20 nor 
a comprehensive exploration of the set of unsolved issues that the site still raises.21 Instead, 
it will discuss some insights that surfaced from my initial multilayered examination of 

et d’architecture libanaises 9 [2005]: 135–8), the area of Kamid el-Loz and the Anti-Lebanon (D. Bonatz, N. Ali, 
and C. Jauss, “Preliminary Remarks on an Archaeological Survey in the Anti-Lebanon,” Bulletin d’archéologie 
et d’architecture libanaises 6 [2002]: 283–307), and the Niha and Ferzol Valleys (P. Newson and R. Young, “The 
Archaeology of Conflict-Damaged Sites: Hosn Niha in the Biqaʿ Valley, Lebanon,” Antiquity 89 [2015]: 449–63). 
Since 2016, Paul Newson has been leading a fieldwork project with the aim of documenting the morphology 
of settlements in the Graeco-Roman period in an area comprised between the Chtaura-Anjar region and the 
headwaters of the River Orontes; see P. Newson, “Archaeological Landscapes of the Bekaa: Past Results and 
Future Prospects,” Berytus 56 (2016): 257–80. It is hoped that this investigation will soon provide new insights 
into the settlement development and land use in this part of the Biqāʿ, with important repercussions for the 
understanding of ʿAnjar’s relation to its immediate hinterland.

16.  B. Finster, “Researches in ʿAnjar, I. Preliminary Report on the Architecture of ʿAnjar,” Bulletin 
d’archéologie et d’architecture libanaises 7 (2003): 209–44; idem, “Vine Ornament and Pomegranates as Palace 
Decoration in ʿ Anjar,” in The Iconography of Islamic Art, Studies in Honour of Robert Hillenbrand, ed. B. O’Kane, 
143‒62 (Edinburgh University Press: Edinburgh, 2005); idem, “Researches in ʿAnjar II. Preliminary Report on 
the Ornaments of ʿAnjar,” Bulletin d’archéologie et d’architecture libanaises 11 (2007): 143–65; idem, “ʿAnjar: 
spätantik oder früislamisch?”

17.  A. Santi, “ʿAnjar in the Shadow of the Church? New Insights on an Umayyad Urban Experiment in the 
Biqāʿ Valley,” Levant 50:2 (2019): 267–80.

18.  See the reference to the so-called “Letters of the Archimandrites” in H. Chehab, “On the Identification of 
ʿAnjar (ʿAyn al-Jarr) as an Umayyad Foundation,” Muqarnas 10 (1993): 42–8, at 43; and Santi, “ʿAnjar,” 12.

19.  The fieldwork was carried out in collaboration with Raffi Gergian, director of the Southern Biqāʿ at 
the Direction Générale des Antiquités du Liban. Regrettably, the social turmoil triggered by anti-government 
protests and the outbreak of the pandemic seriously compromised the planned activities.

20.  A detailed report of the survey, co-authored by Raffi Gergian, is currently in progress. 
21.  Such as the kind of population ʿAnjar was meant to lodge, the reasons behind the disproportion in the 

zoning pattern, and the site’s role in its micro- and macro-regional context.



Al-ʿUṣūr al-Wusṭā 30 (2022)

The Mosque‒Dār al-Imāra Complex at ʿAnjar •  239

the mosque‒dār al-imāra group with the ultimate aim of contributing to a more nuanced 
understanding of the functional role of ʿAnjar in the Marwānid period. Moreover, it will 
try to show that the neat image of a monophasic ʿAnjar we are acquainted with should 
be considered, to some extent at least, the reflection of a deliberate manipulation of the 
archaeological evidence resulted from the heavy, indiscriminate modern interventions at 
the site, rather than a real entity. 

Figure 1. Plan of ʿAnjar (Reworking after Finster 2003)

The Mosque‒Dār al-Imāra Pairing

The mosque‒dār al-imāra pairing in ʿAnjar is located in the south-eastern quadrant of the 
town, at the crossing of the cardo and decumanus (Fig. 1). The presence of this particular 
architectural pairing, which represents the most iconic feature of early Islamic urbanism, 
suggests that ʿAnjar served as, or at least was planned to be, an important administrative 
and ceremonial center for the Marwānid dynasty. In his classification of Umayyad building 



240  •  AilA SAnti

Al-ʿUṣūr al-Wusṭā 30 (2022)

activities, Jere Bacharach had no doubts that the mosque in ʿAnjar was a congregational one 
by virtue of its association with the dār al-imāra, and that the presence of the combination 
rightly qualified the settlement as a madīna,22 showing that the use of the term in early 
Syriac accounts was anything but casual.23 Apart from Bacharach’s comment, however, the 
presence of the pairing in ʿAnjar has never received the attention it deserves. As the only 
example of a mosque-palace complex of the pre-modern Islamic world to have been entirely 
excavated,24 it constitutes an extremely valuable resource for exploring the functional use 
of such a pairing and shedding some light on ritual practices in the Marwānid period. 

22.  J. Bacharach, “Marwanid Umayyad Building Activities: Speculations on Patronage,” Muqarnas 13 (1996): 
27–44, at 35.

23.  An anonymous Syriac chronicle from the year 846 CE (E. W. Brooks, “A Syriac Chronicle of the Year 846,” 
Zeitschrift der Deutschen Morgenländischen Gesellschaft 51 [1897]: 569–88) is one of the only two sources, along 
with the Chronographia by Theophanes the Confessor (Theophanes, Theophanis Chronographia [Bonn: Weber 
1839–41], 1:577), which records the foundation of ʿAnjar in the reign of al-Walīd I. It is noteworthy that the 
Syriac Chronicle describes the foundation as a madīna and uses the toponym ʿIn Gero, which Brooks securely 
identified with the ʿAin Gar mentioned by Yaʿqūbī (Brooks, “Syriac Chronicle,” 581, n. 7). 

24.  Other mosque‒dār al-imāra groups are known to have existed in the main cities of the dār al-Islām but 
were never thoroughly excavated. The earliest extant remains of a dār al-imāra were located by the qibla side 
of the Friday Mosque in Kūfa and have been extensively excavated by the Directorate of Antiquities in Iraq, but 
much uncertainty remains on their dating and their relationship with the mosque, whose early Islamic phases 
have been lost. For a recent discussion on the topic, see A. Santi, “Early Islamic Kūfa in Context: A Chronological 
Reinterpretation of the Palace, with a Note on the Development of the Monumental Language of the Early 
Muslim Élite,” Annali, Sezione Orientale 78 (2018): 69–103; idem, “Reinterpreting the Miṣr: New Insights for 
a Revised Chronological Attribution of the Mosque‒Dār al-Imāra Combination in Kūfa,” in Proceedings of the 
11th ICAANE, vol. 2, ed. O. Adelheid, M. Herles, K. Kaniuth, L. Korn and A. Heidenreich (Harrassowitz Verlag: 
Wiesbaden, 2020), 509–20. An early Islamic mosque-palace group is reported by medieval sources to have existed 
in Baṣra, Kūfa’s twin miṣr, but no systematic archaeological investigations have been carried out at the site of old 
Baṣra yet; see al-Balādhurī, The Origins of the Islamic State, Being a Translation from the Arabic, Accompanied 
with Annotations, Geographic and Historic Notes of the Kitâb Futûḥ al-Buldân of al-Imâm Abu-l ʻAbbâs, Aḥmad 
ibn-Jâbir al-Balâdhuri, trans. F. C. Murgotten (PhD diss., Columbia University, 1924), 60–6; see also L. Massignon, 
“Explication du plan de Baṣra,” in Opera Minora, ed. Y. Moubarac (Beirut: Dar al-Maaref 1963), 61–87. In Wāsiṭ, 
the excavations brought to light the Umayyad mosque‒dār al-imāra complex, but only the first of the two 
buildings was extensively excavated, leaving the extension, form, and plan of the palace utterly unknown; see 
F. Safar, Wāsiṭ: The Sixth Season’s Excavations (Cairo: Imprimerie de l’Institut Français d’Archéologie Orientale, 
1945). Another archaeologically attested case is in Jerusalem, where excavations revealed a palatine complex 
beyond the south and east wall of the Ḥaram al-Sharīf directly connected to al-Aqṣā mosque through a bridge; 
see M. Ben-Dov, In the Shadow of the Temple: The Discovery of Ancient Jerusalem (New York: Harper & Row, 
1985), 324–41, 355–9. However, the results of the excavation still await a comprehensive publication and the date 
of the complex is still the object of debate among scholars. In Iṣṭakhr, Donald Whitcomb suggested that a dār 
al-imāra stood by the qibla side of the mosque based on satellite pictures; see D. Whitcomb, “The City of Istakhr 
and the Marvdasht Plain,” in Akten des VII. Internationalen Kongresses für Iranische Kunst und Archäologie: 
München, 7.-10. September 1976, 363–70 (Berlin: D. Reimer, 1979). Recent excavations carried out at the site, 
however, were unable to confirm this hypothesis; see A. Jaia and L. Ebanista, “Excavations at Istakhr in 2012: The 
Test West of the Site of the Mosque,” in Istakhr (Iran), 2011-2016: Historical and Archaeological Essays, ed. M. V. 
Fontana, 303–44 (Rome: Sapienza Università di Roma, 2018). In Damascus, the Umayyad palace, al-Khaḍrāʾ, was 
reportedly placed in the south-eastern part of the qiblī wall of the mosque, but no remains have been found to 
date; for a discussion of the Qubbat al-Khaḍrāʾ and its relationship with the Umayyad Mosque, see B. Flood, The 
Great Mosque of Damascus: Studies on the Makings of an Umayyad Visual Culture (Leiden: Brill, 2001).



Al-ʿUṣūr al-Wusṭā 30 (2022)

The Mosque‒Dār al-Imāra Complex at ʿAnjar •  241

The two buildings are connected, as is customary, by the qibla side of the mosque, with 
the dār al-imāra located in the direction of prayer. In contrast to other known examples, 
however, they do not share the qiblī wall but are separated by a narrow road. Two matching 
doors, one opening on the northern side of the dār al-imāra, and one in the qiblī wall of the 
mosque, allowed direct transit from a small room in the palace to the central space of the 
prayer hall and were conceivably reserved for the use of the ruler or his representative. 

Despite being the only two buildings in the south-east quadrant—according to Maurice 
Chehab, no remains of any other structures were found elsewhere in this area—the mosque 
and the dār al-imāra are not centrally placed or symmetrically aligned. The presence of 
a peculiar monumental atrium25 preceding the western door of the mosque results in the 
building being shifted eastward by several meters with respect to the palace, which could be 
said to break the symmetry of the Hippodamian grid. Apart from the monumental atrium, 
the mosque follows the classical Damascene prototype, with its long sides oriented towards 
the qibla (which roughly coincides with the geographic south in ʿAnjar), a central courtyard 
flanked by riwāqs on three sides, and a hypostyle prayer hall with columns bearing arches 
(Fig. 2).26 

Figure 2. ʿAnjar: Plan of the mosque (© Santi A.)

25.  In my previous article I interpreted this atrium as an element suggesting that the building was 
originally oriented toward the east. The presence of similar monumental atriums framed by shops and facing a 
colonnaded road is well documented for Byzantine churches in the region, and this, along with other elements, 
induced me to put forth the provocative idea that the mosque of ʿAnjar might have been built over a previous 
basilical church (Santi, “ʿAnjar,” 11). This hypothesis, although intriguing, should be put on hold while awaiting 
further data from new excavations, as the absence of stratigraphic evidence makes it impossible to make more 
substantial and well-founded remarks. 

26.  For a detailed description of the building, see Finster, “Researches in ʿAnjar I,” 229–32.



242  •  AilA SAnti

Al-ʿUṣūr al-Wusṭā 30 (2022)

Unlike the Great Mosque of Damascus, however, the sanctuary here is only two aisles deep 
and, surprisingly, lacks a central transept in front of the miḥrāb. This is a very unusual 
feature, as comparisons with coeval mosques in the Syrian region and beyond27 abundantly 
show that the visibility of the prayer niche was one of the main concerns of planners 
and builders in the Marwānid epoch.28 The lack of such a concern in ʿAnjar is striking, 
as the miḥrāb not only lacks a transept emphasizing its presence, but does not even fall 
in the middle of the adjacent intercolumn. Rather, the central column of the prayer hall 
was placed directly opposite it, deviating from the usual pattern. Remarkably, this odd 
arrangement was not dictated by the need to preserve the centrality of the miḥrāb, since 
the niche is shifted about one meter toward the east of the center of the qiblī wall, with the 
result that it also falls outside the optical axis of the northern entrance, which is correctly 
placed in the exact center of the northern wall.

Such an arrangement strongly suggests that the miḥrāb was not taken into account when 
the general layout of the building was planned, or at least when the mid-row of columns in 
the prayer hall was raised. This implies—as I have pointed out in a previous article29—that 
at least some of the columns in this row pertain to an earlier building phase than that of the 
miḥrāb. Whether this phase predates al-Walīd’s epoch or indicates a sudden change in the 
building plans that occurred in Umayyad times is hard to ascertain. Significantly, however, 
there is no evidence to indicate that the miḥrāb was added to the qiblī wall at a later stage.30 
The small, precisely cut ashlars forming the gradient of the niche fit perfectly into the inner 
face of the qiblī wall, indicating that the wall and the niche were built at the same time 
as part of a unitary project (Fig. 3). Consequently, if our arguments are valid, the central 
columns of the prayer hall existed before the qiblī wall was built. If this was in fact the 
case, it would be extremely unusual, as the qiblī wall was typically the first feature to be 
laid out and built in a mosque, giving the whole building the right orientation toward the  

27.  See, for instance, the Umayyad congregational mosques in ʿAmmān (A. Almagro and P. Jiménez, “The 
Umayyad Mosque of the Citadel of ʿAmmān,” Annual of the Department of Antiquities of Jordan 44 [2000]: 
459–75); Jarash (A. Walmsley and K. Damgaard, “The Umayyad Congregational Mosque of Jarash in Jordan and 
its Relationship to Early Mosques,” Antiquity 79 [2005]: 362–78); Ramla (M. Rosen-Ayalon, “The White Mosque 
of Ramla: Retracing Its History,” Israel Exploration Journal 56 [2006]: 67–83); Ruṣāfa (D. Sack, Resafa IV: Die 
Groẞe Moschee von Resafa [Mainz: P. von Zabern, 1996]); Tiberias (K. Cytryn-Silverman “The Umayyad Mosque 
of Tiberias,” Muqarnas 26 [2009]: 37–61); al-Aqṣā II (R. W. Hamilton, The Structural History of the Aqsa Mosque: 
A Record of Archaeological Gleanings from the Repairs of 1938–1942 [London: Oxford University Press, 1949]).

28.  In this regard, it is important to point out that there are indeed other known cases of mosques with 
miḥrābs misaligned, missing a transept, or obstructed by columns or pillars. Remarkably, however, all of them 
date from later times. Some examples include: the ninth/eleventh-century mosque at Qanat Tepe in Nishapur 
(C. K. Wilkinson, Nishapur: Some Early Islamic Buildings and their Decoration [New York: The Metropolitan 
Museum of Art, 1986] , 264–65); the Ghaznavid mosque at Mt Rāja Gīrā, Udegram (U. Scerrato, “Research on 
the Archaeology and History of Islamic Art in Pakistan: Excavation of the Ghaznavid Mosque on Mt Rāja Gīrā. 
Second Report, 1986,” East and West 36, no. 4 [1986], 496–511); and the sixth/twelfth-century mosque found in 
Segesta (A. Molinari, Segesta II: Il Castello e la Moschea [scavi 1989-1995] [Palermo: Flaccovio, 1997], 95–9. I want 
to thank Maria Vittoria Fontana for acquainting me with these cases.

29.  Santi, “ʿAnjar.”
30.  Finster, “Researches in ʿAnjar I.”



Al-ʿUṣūr al-Wusṭā 30 (2022)

The Mosque‒Dār al-Imāra Complex at ʿAnjar •  243

correct qibla.31 Moreover, all mosques built under Umayyad patronage from the time of 
al-Walīd I onward were planned with a centrally located miḥrāb, which always acted as the 
pivot of the architectural composition and the ultimate focal point of the building.

In an attempt to shed some light on the apparent incongruities of the plan, the new 
survey work targeted the remains of the columns of the prayer hall and the adjacent 
architectural elements. Significantly, this fieldwork showed that some of the observations 
I discussed in my previous article32 need, in fact, to be reassessed. In particular, I argued 
there that the columns in the prayer hall should all be considered as pertaining to an 
older building phase, as I wrongly interpreted the remains of the foundation platforms of 
some of them as plinths. During my recent survey of the building, thanks to the valuable 
help of Raffi Gergian and the DGA working team, it was possible to carry out an accurate 
cleaning of the column bases and the floor of the mosque, which made possible a far clearer 
understanding of their different features and their stratigraphic relationship.

As a first remark, the second survey substantiated what the preliminary study had 
already foreshadowed,33 namely, that the traditional interpretation of the nearly squared 
enclosure surrounding the miḥrāb as a maqṣūra should be reconsidered. Even though the 
loss of the small finds and archaeological stratigraphy prevent us from assigning a precise 
dating to this phase, enough evidence subsists to make it clear that it was extraneous to 
the original plan. The stratigraphic relationship between the structure and the adjoining 
features (in particular the paving slabs of the prayer hall and the three central columns of 
the first row), the building material, and masonry style leave no doubt about it being a later 
addition,34 possibly a subsequent restriction of the prayer hall as attested in other coeval 
mosques in the region such as the Friday mosques in Jarash35 and Ruṣāfa (Fig. 4).36

31.  A. Walmsley, “The Friday Mosque of Early Islamic Jarash in Jordan,” Journal of the David Collection 
1 (2003): 111–31, 124; idem, The Danish-Jordanian Islamic Jarash Project, Season Report 2008 (Copenhagen: 
University of Copenhagen, 2008), 6, 21.

32.  Santi, “ʿAnjar.”
33.  Idem, “ʿAnjar.”
34.  Although Finster claimed that “The quality of the wall and the size of the blocks correspond to the 

masonry of the Great Palace” (Finster, “Researches in ʿ Anjar I,” 231), which made her interpret the structure as a 
maqṣūra, the building technique used for the realization of this structure, lacking any kind of bonding material, 
has nothing in common with that identified in the mosque’s walls, in the adjacent market, or in the walls of the 
palace. The massive, substantial blocks composing this structure resemble in size and finishing some foundation 
blocks found in various rooms of the Great Palace, with the significant difference that they do not function 
as foundations here, since they were clearly laid upon the paving slabs of the prayer hall. This confirms the 
posteriority of the structure to the floor and puts into question the fact that it was part of the original plan, 
suggesting rather that the blocks were Umayyad spolia reused here at a later date. It also casts doubt on the 
interpretation of the enclosure as a maqṣūra as claimed by M. Chehab and Finster. Coeval mosques in the region 
do not provide any fitting comparisons for this kind of enclosure, while, according to written sources, early 
maqṣūras were light wooden structures that would not have survived up to today. The use of such massive 
foundations for a wooden structure is inexplicable and suggests that instead these blocks were set in place to 
sustain a far heavier burden: possibly a roofed wall. 

35.  Damgaard, “The Umayyad Mosque of Jarash,” 193.
36.  Sack, “Resafa IV,” 44–5.



244  •  AilA SAnti

Al-ʿUṣūr al-Wusṭā 30 (2022)

Figure 3. ʿAnjar: the miḥrāb (Photo by the author, 2017)

Figure 4. ʿAnjar: the so-called maqṣūra (Photo by the author, 2017)



Al-ʿUṣūr al-Wusṭā 30 (2022)

The Mosque‒Dār al-Imāra Complex at ʿAnjar •  245

The analysis of the remains of the middle colonnade of the prayer hall raised further 
important questions about the alleged monophasic nature of the mosque. There, it was 
possible to identify three different building techniques in use for the construction of the 
nine columns of the arcade, which allowed their division into three main groups. The first 
one, comprising the first three columns starting from the east, is characterized by the 
presence of a nearly square foundation platform and a smaller square plinth carved from 
the same block of stone, which probably used to bear a baseless column shaft (Fig. 5). 

Figure 5. ʿAnjar, inner arcade of the prayer hall of the mosque:  
remains of the 3rd column from the east (Photo by the author, 2020)

The second group, encompassing the last two supports to the west, features reused Attic 
bases with square plinths bonded to square foundation platforms by a thick layer of mortar 
(Fig. 6).37 Finally, the three central columns of the row are distinguished by the fact that 
their Attic bases, square plinths, and foundation platforms were all carved from one piece 
(Fig. 7). Some targeted cleaning made it possible to ascertain that the dressed faces of 
the square plinths seamlessly turn into coarse rocky foundations at a depth of about 10 
centimeters under the floor level. Unlike the other foundation platforms, which feature a 
roughly levelled surface bearing visible signs of dressing, the lower parts of these supports, 
whose extension in width and depth cannot be defined due to the presence of the paving 
slabs, look like rock in a natural state (Fig. 8). Remarkably, the supports of this type are set 
at a lower level than the others in the row, being the plinths of the bases partially covered 
by the paving slabs—whereas in all the other cases the foundation platforms protrude 
slightly above the floor level.

37.  This type of mortar, white/reddish in color and rather clear, has not been found elsewhere in the 
mosque, where the use of a grey mortar with big black inclusions prevails. This may indicate that these columns 
also date to a different building phase than the rest of the mosque, even though no further evidence has been 
found to support this hypothesis. 



246  •  AilA SAnti

Al-ʿUṣūr al-Wusṭā 30 (2022)

Figure 8. ʿAnjar, inner arcade of the prayer hall of the mosque:  
remains of the 5th column from the east, detail (Photo by the author, 2020)

One might argue that the different building techniques employed in the making of the 
columns were responses to different structural exigencies in different parts of the building, 
such as the need to sustain a heavier load in correspondence to the monolithic bases and 
foundations of the three central columns. An analysis of the possible roofing solutions for 

Figure 6. ʿAnjar, inner arcade of the 
prayer hall of the mosque: remains  

of the 9th column from the east  
(Photo by the author, 2020)

Figure 7. ʿAnjar, inner arcade of the 
prayer hall of the mosque: remains  

of the 5th column from the east  
(Photo by the author, 2020)



Al-ʿUṣūr al-Wusṭā 30 (2022)

The Mosque‒Dār al-Imāra Complex at ʿAnjar •  247

the prayer hall, however, disproved this hypothesis. The emplacement of the columns and 
the regular width of the intercolumns and naves suggest that the mosque in ʿAnjar was 
covered by a tiled, gabled roof carried on wooden beams of the same kind as those of the 
mosques of Jarash and Ruṣāfa, and that the weight of the roof was regular along the whole 
length of the colonnade (Fig. 9). Since there is no structural justification for the use of 
different building techniques, we should consider the possibility that they might instead 
have some chronological significance.

Figure 9. Isometric views of the Umayyad mosques of  
Ruṣāfa (left, after Sack 1996) and Jerash (right, after Damgaard, 2011)

As for the first three and last two columns of the row, the presence of substantial 
foundation platforms of probable Umayyad making38 suggests that we should consider the 
bases associated with them—of the standard Attic type carved together with a square plinth, 
typical of the Syrian region in Roman and Byzantine time39—as being Roman or late antique 

38.  The finish of the blocks composing the foundation platforms of these columns is not dissimilar from that 
of the Umayyad T-pilasters found in the saḥn, which might suggest the same dating for these features. 

39.  The Attic-type base developed in Greece in the fifth century BCE and became a standard type of base in 
the eastern Roman provinces, in Syria, Phoenicia and Palestine (L.S. Meritt, “The Geographical Distribution of 
Greek and Roman Ionic Bases,” Hesperia 38 (1969): 186–204, at 196, n. 41–2). The Attic bases cut from a single 
block together with plinths were introduced in the region only in the first century BCE, presumably due to 
Roman influence, and continued to be used in Christian buildings up to at least the sixth century CE; see M. 
Burdajewicz, “Some Remarks on the Architectural Decoration of the North-West Church at Hippos (Susita),” in 
Ex Oriente Lux: Studies in Honour of Jolanta Młynarczyk, ed. K. Jakubiak and A. Łajtar (Warsaw: University of 
Warsaw Press, 2020). 



248  •  AilA SAnti

Al-ʿUṣūr al-Wusṭā 30 (2022)

displaced spolia that were reused in Umayyad times in the construction of the mosque. In 
contrast, the technique attested for the three central columns, in which the foundation 
platforms appear to be seamlessly merged with the upper base-with-plinth element, doesn’t 
find—to the best of my knowledge—any fitting parallel in antique, late-antique, or coeval 
buildings in the region. Nonetheless, since the bases-with-plinth represent perfectly fine 
examples of the Attic order in its eastern version, it is most likely that these composite 
pedestals also date to the Roman or Byzantine period. Nevertheless, the depth and 
substantialness of their foundations, and their being set at a significantly lower level than 
the others in the row, may be an indication that they stand in their original position rather 
than being reused spolia. In fact, this idea gains plausibility when considered in the light of 
the planimetric incongruencies of the building. As pointed out earlier, the major problem 
with the mosque of ʿAnjar is the lack of any device emphasizing the visibility of the miḥrāb 
as a result of the eccentric placement of the column facing the niche. This occurrence is 
very hard to justify unless we consider the possibility that the planning of the qiblī wall was 
constrained by the pre-existence of the central columns. 

In fact, this possibility seems likely even when we consider the relationship of the 
mosque with the adjacent dār al-imāra. An overall view of the complex (Fig. 10) reveals that 
the eccentric positioning of the miḥrāb was in part determined by the need to create an 
entrance in the qiblī wall to match the door on the northern wall of the palace. This allowed 
the caliph or his representative to access the mosque from the dār al-imāra through a direct 
and preferential access, the so-called “bāb al-imām.” This would imply that the complex is 
not the outcome of a unitary plan, but that the palace was probably laid out and built prior 
to the mosque, insofar as the planning and building of the qiblī wall were determined by the 
need to fit the arrangement of the palace. 

If that was the case, however, and no pre-existing elements stood on the site of the 
mosque, the visibility and centrality of the miḥrāb could have been preserved by building 
the columns of the prayer hall in such a way that the niche would have fallen in the center 
of the adjacent intercolumn. The fact that this was not done and that, in addition, the 
easternmost of the three central columns falls just in the way of the second door in the 
qiblī wall (whose function will be discussed below) constitutes a further indication of the 
potential pre-existence of the central part of the colonnade before the building of the 
mosque in its final form. Even if one accepts the pre-existence of the columns to the qiblī 
wall, however, the visibility of the miḥrāb could have been preserved by simply placing it 
slightly to the east, so that it would have been at the center of the adjacent intercolumn. 
This solution was probably rejected to maintain a semblance of centrality of the niche in 
the qiblī wall.

It could be asked at this point why, then, was the palace not constructed with due 
attention to the pre-existing columns of the building to its north? In fact, this could have 
been the result of uncoordinated planning of the two individual buildings that made up the 
complex, along with the possible presence of further pre-existences at the site of the palace, 
which could have conditioned and constrained its planning and construction to some 
extent. In conclusion, what can be said with certainty is that, after a close and attentive 
look, the mosque‒dār al-imāra complex in ʿAnjar appears more like the result of a clumsy 



Al-ʿUṣūr al-Wusṭā 30 (2022)

The Mosque‒Dār al-Imāra Complex at ʿAnjar •  249

attempt to harmonize a set of disjointed and possibly non-contemporary structures than 
the pristine outcome of a unitary project. 

Figure 10. ʿAnjar: mosque-dār al-imāra combination (© Santi A.)

Ceremonial Aspects of the Mosque‒Dār al-Imāra Complex in ʿAnjar 

The analysis just carried out allows us to acknowledge with relative certainty that the 
mosque gained its current form as a result of the conditioning of pre-existent structures, 
and to facilitate the transit of the caliph or his representative from a tiny room in the 
dār al-imāra into the central space of the prayer hall. The creation of this special path 
was considered important to the extent that the visibility and centrality of the miḥrāb—
normally the most important thing in a Marwānid mosque—was sacrificed. The emphasis 
given to the bāb al-imām to the detriment of the prayer niche suggests that the former 
retained a particular ceremonial significance.

In the search for a functional interpretation for this ensemble, it should be pointed out 
that the arrangement of the qiblī wall in the mosque of ʿAnjar and its relationship with 
the adjacent dār al-imāra closely recalls the layout of the Tulunid mosque–palace group in 



250  •  AilA SAnti

Al-ʿUṣūr al-Wusṭā 30 (2022)

al-Qaṭāʾiʿ, present-day Cairo (Fig. 11).40 This complex retains a special significance as it is the 
only case of a mosque‒dār al-imāra combination whose ceremonial use is described in detail 
in the written sources. Even though nothing is left of the Tulunid dār al-imāra,41 we know 
from historic accounts that it used to abut the qibla side of the mosque and gave direct 
access to the prayer hall through two doorways opened in the qiblī wall, one to the right 
and one to the left of the miḥrāb. 

Figure 11. Tulunid dār al-imāra-mosque complex at al-Qaṭāʾi (Swelim 2015)

Al-Balawī and al-Maqrīzī narrate that, on Fridays, Aḥmad b. Ṭulūn used to reach the complex 
from his residential palace in al-Maydān through the so-called shāriʿ aʿzam (“Greatest 
Street”).42 Once arrived in the dār al-imāra, he would have performed the ritual ablutions, 
changed his garments and robes, and perfumed himself with incense. When ready, he would 

40.  For a recent study of the complex and its ceremonial use, see T. Swelim, Ibn Tulun: His Lost City and 
Great Mosque (Cairo: AUC Press, 2015). 

41.  Some minor rooms on the qibla side of the present mosque have been interpreted as part of the lost dār 
al-imāra; see in this regard A. F. Sayyid, La capitale de l’Égypte jusqu’à l’époque fatimide, al-Qāhira et al-Fusṭāṭ 
(Stuttgart: Franz Steiner Verlag, 1998), 48.

42.  Swelim, Ibn Tulun, 37.



Al-ʿUṣūr al-Wusṭā 30 (2022)

The Mosque‒Dār al-Imāra Complex at ʿAnjar •  251

have entered the prayer hall through the door on the left side of the main miḥrāb to attend 
the Friday prayer in front of the congregation of subjects.43 

Although over a century and a change of dynasty separate the foundation of ʿAnjar and 
the building of the Mosque-Palace group in al-Qaṭāʾiʿ, the latter complex, being shaped 
upon a Samarran model,44 descends directly from the tradition of mosque‒dār al-imāra 
pairings of the Umayyad period.45 This is also confirmed by the identical arrangement of 
the ceremonial spaces in the two complexes, which suggests that the group of rooms in the 
northern wing of the dār al-imāra of ʿAnjar, with their proximity and direct connection to 
the sanctuary of the mosque, had a function associated with the ritual preparation of the 
ruler for the Friday prayer as depicted in the  report about Ibn Ṭulūn. An undocumented 
piece of evidence found during my last survey of the palace in ʿAnjar is worth mentioning 
in this connection. This comprises a rectangular tank excavated below the floor level of the 
south-eastern room flanking the northern audience hall, covered and bordered by finely cut 
and dressed blocks of limestone (Fig. 12). 

Figure 12. ʿAnjar: dār al-imāra. Remains of a tank (Photo by the author, 2017)

43.  Idem, Ibn Tulun, 47, 127.
44.  As observed by Matthew Gordon, in joining a downsized dār al-imāra to the Friday Mosque, Ibn 

Ṭūlūn was following long-established early Islamic practice; at least two and possibly three examples of this 
arrangement occurred in Samarra. See M. S. Gordon, “Ibn Ṭūlūn, al-Qaṭāʾiʿ and the Legacy of Samarra,” Beiträge 
zur Islamischen Kunst und Archäologie 4 (2014): 63–77, at 69.

45.  That said, during the Abbasid period, the dār al-imāra abutting the mosque gradually lost its residential 
functions and remained only as an administrative and/or ceremonial appendix to the Friday Mosque. For a 
detailed assessment of the development of the mosque‒dār al-imāra combination in the Abbasid period, see 
F. Duva, Persistenza e sviluppo del dispositivo ‘moschea del venerdì–dār al-ʾimāra’ nell’area iranica in periodo 
proto-abbasid (PhD diss., Sapienza Università di Roma, 2021).



252  •  AilA SAnti

Al-ʿUṣūr al-Wusṭā 30 (2022)

Traces of weathering and the remains of thick layers of hydraulic mortar suggest that it 
could have served as a water container, possibly as part of a structure intended for the 
performance of exclusive ritual ablutions. 

As for the door on the right hand side of the miḥrāb, we know that in the mosque of 
Ibn Ṭulūn it was allocated for the use of the khaṭīb, the authority in charge of delivering 
the sermon on Fridays, and was connected to an adjacent room in the dār al-imāra for the 
exclusive use of the preacher.46 On this note, it is key to point out that the khaṭīb became 
a government-appointed position only under the Abbasids, and in Marwānid times, when 
the mosque‒dār al-imāra complex of ʿAnjar was designed, the office was still an exclusive 
prerogative of the caliph or his appointed governor.47 This implied that, when attending 
the Friday prayer, the Marwānid leader had to display tokens of honor and authority 
traditionally associated with the figure of the khaṭīb in pre-and early Islamic times,48 and 
pronounce the khuṭba, in this period an address of political nature rather than a religious 
sermon. Based on that, we must suppose that the function of the ancillary rooms of the 
northern audience hall in the dār al-imāra of ʿAnjar was not limited to ablutions, change 
of clothes, and fumigations, but also to the preparation of the khuṭba and storage of the 
insignia of the khaṭīb’s office, in particular, some kinds of elongated pointed objects—such 
as the lance (ʿanaza), staff (qaḍīb) or bow (makhāṣir)—which the khaṭīb held in his right 
hand during the khuṭba and which, due to their connection to imagery associated with 
the Prophet Muḥammad,49 represented the ultimate symbols of caliphal authority in this 
period.50 

In fact, written sources abundantly attest51 that the entry of the amīr al-muʾminīn or his 
representative into the congregational space was strictly regulated, in the late Umayyad 
period, by a precise sumptuary code due to the need to exalt the religious component of the 

46.  Swelim, Ibn Tulun, 127.
47.  “It is quite in keeping with the nature of early Islam and with that of the Arab khaṭīb that the ruler 

himself was spokesman and that he not only made edifying speeches from the minbar as khaṭīb but also issued 
orders, made decisions and pronounced his views on political questions and particularly questions of general 
interest. This was the case under the first four caliphs and the Umayyads, and the governors appointed by them 
also acted as khuṭabāʾ” (J. Pedersen, “Khaṭīb,” in The Encyclopaedia of Islam, 2nd ed., ed. P. Bearman et al.,  4:1, 
109–11 (Leiden: Brill, 1954–2009), 1,110.

48.  Pedersen, “Khaṭīb.”
49.  According to the tradition, the Prophet came forward as a khaṭīb after the conquest of Mecca (Pedersen, 

“Khaṭīb,” 1,110).
50.  For the association of the staff with the ʿ anaza of the Prophet, see G. C. Miles, “ʿAnaza,” in The Encyclopaedia 

of Islam, 2nd ed., ed. P. Bearman et al., 1:482 (Leiden: Brill, 1954–2009); idem, “Miḥrāb and ʿAnazah: A Study 
in Islamic Iconography,” in Early Islamic Art and Architecture, ed. J. M. Bloom, 149–66 (London: Routledge, 
2002); L. Treadwell, “‘Mihrab and ʿAnaza’ or ‘Sacrum and Spear’? A Reconsideration of an Early Marwanid Silver 
Drachm,” Muqarnas 22 (2005): 1–28; H. Munt, The Holy City of Medina, Sacred Space in Early Islamic Arabia (New 
York: Cambridge University Press, 2014), 104.

51.  See for instance al-Masʿūdī, Les prairies d’or, ed. and trans. C. Barbier de Maynard (Paris: Imprimerie 
Impériale, 1869), 5:402–3; al-Ṭabarī, The History of al-Ṭabarī (Taʾrīkh al-rusul waʾl mulūk), vol. 24, The Empire in 
Transition, trans. D. S. Powers (Albany, NY: State University of New York Press, 1989), 62.



Al-ʿUṣūr al-Wusṭā 30 (2022)

The Mosque‒Dār al-Imāra Complex at ʿAnjar •  253

caliphal office.52 As Oleg Grabar demonstrated in his PhD dissertation, the use of particular 
types of headgear drawn from the Sasanian tradition, such as the qalansuwa ṭawīla53 and 
the tāj,54 which were associated with the secular, royal aspect of Umayyad authority, were 
categorically substituted, in the mosque, by the turban (ʿimāma), the headgear of the 
Prophet.55 This, along with the wearing of the “caliphal clothes” (thiyāb al-khilāfa),56 and 
the performance of the ritual ablutions, allowed the caliph to enter a state of ceremonial 
purity57 and complete his ritual transformation into the imām and heir to Muḥammad’s 
religious authority. The part of the dār al-imāra adjacent to the qiblī wall would have 
therefore acted as a liminal space for the ruler and the setting of this ritual transformation. 
Within the context of this functional interpretation, the location of the bāb al-imām, which 
in ʿAnjar falls within the optical axis of the northern door of the mosque to the detriment 
of the visibility of the miḥrāb, was probably an intentional device designed to emphasize 
the spectacular entrance of the ruler into the mosque, adorned with the insignia of his 
multifaceted office. This would have produced in the onlookers an affectively powerful and 
semiotically meaningful experience. 

There is one particular piece of architectural decoration that was retrieved during 
Chehab’s excavation and which has been, unfortunately, very little studied, that may be of 
some interest in this context. It is a capital that is currently part of the arcade close to the 
north gate but which was probably located elsewhere originally58 which portrays, inside a 
roundel surrounded by acanthus leaves, a frontal figure holding two objects in both hands 
(Figure 13). 

52.  O. Grabar, Ceremonial and Art at the Umayyad Court (PhD diss., Princeton University, 1955).
53.  High, miter-like Persian headgear which was was a symbol of royal authority in the Umayyad period. For 

an overview of the Umayyad attestations of the qalansuwa ṭawīla, both in art and in the written sources, see R. 
Ettinghausen, From Byzantium to Sasanian Iran and the Islamic World: Three Modes of Artistic Influence (Brill: 
Leiden, 1972): 28–34 and related plates.

54.  A term that refers to the crown in a general sense.
55.  Significantly, Muḥammad was referred to by the later tradition as “ṣāḥib al-ʿimāma,” “the one who wears 

the turban”; see Y. K. Stillman, “Libās,” in The Encyclopaedia of Islam, 2nd ed., ed. P. Bearman et al., 5:732–42 
(Leiden: Brill, 1954–2009), at 734. According to a saying of al-Ṭabarī, the turban was considered “the crown of 
the Arabs” in the early Islamic period; see A. Marsham, Rituals of Islamic Monarchy: Accession and Succession 
in the First Muslim Empire (Edinburgh: Edinburgh University Press, 2009), 140. As rightly pointed out by Grabar: 
“The turban was the headgear of the caliph as imām, that is, as successor of the Prophet performing a religious 
ceremony” (Grabar, Ceremonial and Art, 59).

56.  Y. K. Stillman, Arab Dress: A Short History from the Dawn of Islam to Modern Times, 2nd ed. (Leiden: 
Brill, 2003), 34; Grabar, Ceremonial and Art, 63.

57.  Many early traditions demonstrate the existence, since the early Islamic period, of definite rules 
regulating the attire in connection to the state of ritual purity or impurity of a person. Remarkably, some type 
of garments including the ʿimāma were specifically forbidden for men in a state of ritual impurity (iḥrām); see 
Stillman, Arab Dress, 23. 

58.  Many architectural elements were moved and arbitrarily relocated during the restoration works directed 
by Houti Kalayan, as the unpublished photographic documentation of the DGA abundantly demonstrates. 



254  •  AilA SAnti

Al-ʿUṣūr al-Wusṭā 30 (2022)

Figure 13. Capital with relief featuring a possible caliphal portrait  
(Photo by the author, 2020)

The poor state of conservation of the relief notwithstanding, the head of the figure 
seems to be characterized by long, coiffured hair and what looks like voluminous but short 
headgear, while the pointed object wielded in the left hand seems to be identifiable as a 
spear. Analogous Marwānid long-haired figures in frontal representation wielding weapons 
and wearing diadems/headgear have survived in different types of media59 and are generally 
 recognized to pertain to royal or princely iconography.60 If we accept the possibility of this 

59.  Notably, on coins of the “standing caliph” type issued during the first phase of ʿAbd al-Malik’s 
monetary reform (S. Heidemann, “The Evolving Representation of the Early Islamic Empire and its Religion 
on Coin Imagery,” in The Qurʾan in Context: Historical and Literary Investigations into the Qurʿanic Milieu, 
ed. A. Neuwirth, N. Sinai, and M. Marx, 149–195 [Leiden: Brill 2010], 175–6, figs 21–4), in two round stucco 
sculptures from Qaṣr al-Ḥayr al-Gharbī (D. Schlumberger, Qasr el-Heir el-Gharbi [Paris: P. Geuthner, 1986], 15, 
22) and Khirbat al-Mafjar (R. W. Hamilton, Khirbat al-Mafjar: An Arabian Mansion in the Jordan Valley [Oxford: 
Clarendon Press, 1959], 237–9, pl. XXXVI 3 and 7), by some interpreted as portraits of the caliph Hishām b. ʿAbd 
al-Malik; and on less known objects such as a series of ivory carvings from al-Ḥumayma in southern Jordan (R. 
M. Foote, “An Abbasid Residence at al-Humayma,” in Byzantium and Islam, Age of Transition, 7th–9th Century, 
ed. H. C. Evans and B. Ratliff, 221–2 [New York: Metropolitan Museum of Art, 2012]), a newly-discovered stucco 
panel from Building E at Qaṣr al-Ḥayr al-Gharbī (D. Genequand, “Les décors en stuc du bâtiment E à Qaṣr al-Ḥayr 
al-Sharqī,” Syria 88 [2011]: 351–78), and an ivory plaque kept at the Walters Art Gallery in Baltimore (acc. no. 
71.62, see M. Di Cesare, “A Note on an Umayyad Carved Ivory Plaque Kept at the Walters Art Gallery,” Vicino 
Oriente 21 [2017]: 197–210). For a recent discussion of the topic, see D. Genequand, “Two Possible Caliphal 
Representations from Qaṣr al-Ḥayr al-Sharqı̄ and Their Implication for the History of the Site,” in Power, 
Patronage, and Memory in Early Islam: Perspectives on Umayyad Elites, ed. A. George and A. Marsham, 147–74 
(New York: Oxford University Press, 2017).

60.  Even though the lack of accompanying inscriptions makes it impossible to securely identify these 
portraits as caliphal representations—with the exception of the coins of the so-called “standing caliph” type—
their connection with the spheres of authority, high status, and often even kingship is generally accepted as 
factual. 



Al-ʿUṣūr al-Wusṭā 30 (2022)

The Mosque‒Dār al-Imāra Complex at ʿAnjar •  255

being a caliphal portrait—a belief shared by both Finster and the late Hafez Chehab61—then 
it could be speculated that the attributes of the figure might be a reference to the prophetic 
regalia abovementioned: in particular, the ʿanaza62 and the ʿimāma of Muḥammad. If that 
were in fact the case, the image portrayed was probably not very different from that seen 
on Fridays in the mosque of ʿAnjar and could be a further indication of a caliphal presence 
in the foundation of the site.

Returning to the mosque‒dār al-imāra combination, since in Marwānid times the bāb 
al-khaṭīb was not distinguished from the bāb al-imām, the door on the eastern side of 
the miḥrāb in the mosque of ʿAnjar needs an alternative interpretation. We have already 
discussed the reasons why the enclosure around the miḥrāb should be considered a later 
addition rather than an Umayyad maqṣūra. This implies that, originally, the bāb al-imām 
and the door east of the miḥrāb provided equal access to the central space of the prayer 
hall, which perhaps was enclosed in a more conventional wooden maqṣūra. The perfect 
alignment of the eastern qibla door with the termination of the north-eastern corridor of 
the palace suggests that the latter was, in turn, provided with a matching doorway, in the 
same way the small service room adjoining the apse of the northern hall has a door facing 
the bāb al-imām. In fact, the section of the northern wall of the palace aligned with the 
eastern qibla door is preserved only at the level of the foundation ashlars and, since its state 
has remained unaltered since the excavations,63 the presence of a door in this spot can be 
confidently proposed (Fig. 14). This arrangement suggests that this entrance was used by a 
special category of subjects to enter the central space of the prayer hall directly from the 
palace and attend the Friday service in close proximity to the ruler. Indeed, we know from 
the sources that the political leader never remained entirely alone in the maqṣūra: the 
ḥaras, or caliphal guards, always stood behind him, along with members of the Umayyad 
family or other important individuals who were granted the privilege of being introduced 
into the royal enclosure.64

61.  The piece has been very briefly discussed by Finster, who interpreted it as a portrait of the Byzantine 
emperor (Finster, “Palace Decoration in ʿAnjar,” 155) or a portrait of the caliph with the attributes of the 
Byzantine ruler to symbolize his universal power (Finster, “ʿAnjar: spätantik,” 233). It is not clear why Finster 
interpreted the image in these ways, and which element she specifically associated with a Byzantine royal 
iconography. The image has been republished in the recent posthumous publication of Hafez Chehab’s PhD 
dissertation, accompanied by a detailed description. Chehab securely identified it as a caliphal portrait due 
to the presence of the ʿanaza, “la lance qui était l’un des symbols de la puissance califienne.” See H. Chehab, 
Les ruines d’Anjar, revu et annoté par G. Homsy-Gottwalles (London: Lebanese British Friends of the National 
Museum, 2017–18).

62.  The spear in this case could also represent a mere implement of war and/or “an instrument of God’s 
wrath against the nonbelievers,” as Luke Treadwell has interpreted its iconography in the famous series of 
drachms of the so-called miḥrāb and ʿanaza types (Treadwell, “Mihrab and ʿAnaza”).

63.  This was verified thanks to unpublished pictures of the excavations made available to me courtesy of 
the DGA.

64.  J. Sauvaget, La mosquée omeyyade de Médine (Paris: Institut Français de Damas, 1947), 149–51.



256  •  AilA SAnti

Al-ʿUṣūr al-Wusṭā 30 (2022)

Figure 14. ʿAnjar, eastern qibla door of the mosque aligned  
with the end of a corridor of the adjacent dār al-imāra  

(Photo by the author, 2020)

It can be suggested that in ʿAnjar the second qiblī doorway was intended as a device 
to allow the guards to enter the mosque simultaneously with the ruler without visually 
affecting the spectacular impact of his public appearance. On the other hand, using the 
second qiblī door was probably also regarded as a token of the highest honor reserved for 
special individuals or families. An interesting precedent exists in this regard in the Prophet’s 
Mosque in Medina, where numerous sources report the existence of a small door (khawkha) 
to the east of the miḥrāb, which was intended for the exclusive use of the descendants of 
ʿUmar b. al-Khaṭṭāb (Fig. 15).65 This correspondence reveals that the mosque-palace group 
in ʿAnjar was shaped in accordance with a most celebrated architectural model and that the 
ritual dimension was an intrinsic, essential aspect of the site. In conclusion, whether the 
mosque‒dār al-imāra group in ʿAnjar ever functioned as a ceremonial venue or not, what is 
evident is that the complex was designed to be, or to symbolically evoke at least, a setup for 
splendid ceremonials. 

As a final remark, I would like to propose a connection between ʿ Anjar’s vibrant ceremonial 
character and the alleged presence of pre-existing structures at the site which were 
hypothetically identified in this paper as part of the columns of the prayer hall, but whose 
extent and diffusion throughout the site area could have possibly been more substantial.  

65.  See A. Santi, “The role of Madīna in the Emergence of the Mosque‒Dār al-Imāra Combination: A 
Preliminary Note,” Vicino Oriente XXI (2017): 211–23, at 218.



Al-ʿUṣūr al-Wusṭā 30 (2022)

The Mosque‒Dār al-Imāra Complex at ʿAnjar •  257

Figure 15. Medina, the mosque-dār al-imāra complex  
as reconstructed by the author (© Santi A.)

Although the existence of pre-Islamic vestiges at the site has been denied by the 
excavators,66 the lack of documentation, the invasive rebuilding works, the loss of the finds, 
and the destruction of part of the archaeological evidence suggest that it would be wise to 
be cautious in advancing any categoric assumption in this regard. In fact, both Grabar67 and 
Hafez Chehab68 voiced doubts over whether ʿAnjar—whilst undoubtedly being an Umayyad 
construction project—should be definitely considered an Umayyad foundation ex nihilo. In 

66.  As I pointed out in my previous article, although M. Chehab in his paper asserted the Umayyad foundation 
of ʿAnjar ex-nihilo as a matter of fact, he also alluded indirectly to the pre-existence of an important Roman 
settlement at the site (Chehab, “Umayyad Palace,” 17). Moreover, his comment that the excavators at the site 
“often struck virgin soil” (Chehab, “Umayyad Palace,” 19), raises the question of whether they may occasionally 
have struck pre-Umayyad remains (see Santi, “ʿAnjar,” 6).

67.  O. Grabar, The Formation of Islamic Art: Revised and Enlarged Edition (New Haven, CT: Yale University 
Press, 1987), 168; idem, “Umayyad Palaces Reconsidered,” Ars Orientalis 23 (1993): 93–108.

68.  H. Chehab, “Les palais omeyyades d’Anjar, résidences princières d’été,” Archéologia 87 (1975): 18–25; 
idem, “Identification,” 42–8.



258  •  AilA SAnti

Al-ʿUṣūr al-Wusṭā 30 (2022)

this regard, Chehab mentioned an interesting piece of evidence found in a document written 
in 569 as a declaration of Miaphysite faith and undersigned by 137 archimandrites from 
Provincia Arabia and Phoenicia Libanensis69 known as “the letter of the archimandrites.” 
In particular, the scholar drew attention to one of the signatures, which reads: “I, Paulos, 
deacon and head of the monastery of ʿYNGDʾ (or ʿYNGRʾ), have written by the hand of ḤLPY 
of the monastery of the God-loving ḤLPY of the village of DWRBYL.”70 Interestingly the 
first toponym, which records the name of the place where Paulos’ monastery was located, 
has among its most probable readings “ʿAyn Jara.” As noted by Nöldeke,71 however, the 
name may refer to the site of ʿAyn Jara in Jordan, west of Jarash, which, in pre-Islamic 
times, was part of the ancient Provincia Arabia. This notwithstanding, Chehab pointed out 
that the second recorded toponym, the one where the epistle was signed, DWRBYL, while 
it has no correspondents in the area of Jarash, very much recalls the modern toponym of 
Turbul, a Christian rural agglomeration some two hours walk north of ʿAnjar. If we accept 
this identification, we may hypothesize that the toponym ʿAyn Jara in this part of the Biqāʿ 
predated the Islamic conquest and might have been associated with a monastic precinct in 
the area.72

This hypothesis, although extremely tentative given the current fragmented state 
of archaeological knowledge of the site and its surroundings, appears to be especially 
intriguing in terms of the selection of this location by the Umayyads for the establishment 
of a settlement. Indeed, the area of ʿAnjar happens to be along the itinerary conceivably 
followed by ʿAbd al-Malik and his court every year when, with the coming of the summer 
heat, they moved from the monastery of Dayr Murrān on the western outskirts of Damascus73 
(where the caliph used to spend part of the spring) to Baʿalbaak. Located on the ancient 
caravan road leading to Roman Heliopolis, on the only viable path through the marshy lake 
of the Biqāʿ, and close to the karstic perennial springs of the Anti-Lebanon, the site of ʿAnjar 
must have been an ideal place for the caliphal caravan to break the one-day long journey to 
Baʿalbaak, and even more so in the hypothetical presence of a monastic installation.74 

69.  The Roman province that encompassed the eastern Biqāʿ including the area of ʿAnjar.
70.  F. Millar, “Christian Monasticism in Roman Arabia at the Birth of Mahomet,” Semitica et Classica 2 (2009): 

97–115, at 111.
71.  T. Nödelke, “Zur Topographie und Geschichte des Damascenischen Gebietes und der Haurângegend,” 

Zeitschrift der Deutschen Morgenländischen Gesellschaft 29/34 (1876): 419–44, at 441; see also Chehab, 
“Identification,” 43.

72.  I already advanced this hypothesis in my previous article; see Santi, “ʿAnjar,” 277–8.
73.  H. Kennedy, The Prophet and the Age of the Caliphates: The Islamic Near East from the Sixth to the 

Eleventh Century (London: Routledge, 2015), 110.
74.  Interestingly, Ghassanid Monophysite monastic installations were often strategically located at 

crossroads, near well-traveled routes or perennial water sources. This is strictly linked to their role as places of 
proselytism and territorial (and hence political) control in the pre-Islamic period, which underlies the pattern of 
re-use that linked Roman forts, monasteries, and Umayyad settlements in the shift between Late Antiquity and 
the Islamic era. This phenomenon, its roots, and its ramifications, have been thoroughly examined by Ignacio 
Arce in “Romans, Ghassanids and Umayyads and the Transformation of the Limes Arabicus: From Coercive 
and Deterrent Diplomacy Towards Religious Proselytism and Political Clientelarism,” in La Transgiordania nei 
secoli XII-XIII e le ‘frontiere’ del Mediterraneo medievale, ed. G. Vannini and M. Nucciotti, 55–74 (Oxford: BAR 



Al-ʿUṣūr al-Wusṭā 30 (2022)

The Mosque‒Dār al-Imāra Complex at ʿAnjar •  259

The tight relationship between the Umayyad court and monasteries in Greater Syria 
is proved by archaeology, as we know of at least seven Umayyad settlements built above 
remains of Christian monasteries, including al-Ruṣāfa, Qaṣr al-Hayr al-Gharbī, and 
Qaṣr al-Ḥallābāt.75 This relationship has been variously investigated and explained in 
symmetrical terms.76 Monastic installations and Umayyad foundations in al-Shām shared 
a set of practical, spiritual, aesthetic, and political functions linked to diplomacy and the 
exploitation of the land. The role of Umayyad complexes in making the dynasty’s presence 
felt in rural areas, and facilitating surveillance of the tribes on which their authority relied, 
evokes parallels with the Ghassanid and Lakhmid use of monasteries, which often served as 
a backdrop for exhibitions of influence and control.77 This, along with the ideal ecological 
setting and perfect strategic position of ʿAnjar,78 could have played a role in the choice of 
the site for the foundation of a new madīna with ritualistic and ceremonial overtones.

Conclusions

This paper has tried to highlight how the idea of a monophasic ʿAnjar entrenched 
in literature needs to be reconsidered, and how the site can still provide, if thoroughly 
examined, valuable glimpses into the Marwānid liturgy of state, court culture, and territorial 
policy. In fact, the particular configuration of the mosque‒dār al-imāra pairing and the 
vibrant ceremonial character that emerged from its analysis makes of ʿAnjar a quite unique 
case among the madīnas of Bilād al-Shām.79

Although fitting the categorization of madīnas as defined by Denis Genequand80 in terms 
of its size, orthogonal plan, fortified enclosure walls, economic, religious, political, and 

Publishers), 68. 
75.  But also, Qaṣr Burquʿ; al-Fudayn; Qaṣr al-Baʾij; Dayr al-Kahf; and Dayr al-Qinn. See E. K. Fowden, “Christian 

Monasteries and Umayyad Residences in Late Antique Syria,” Antigüedad y Cristianismo 21 (2004): 565–81.
76.  See in particular Arce, “Romans, Ghassanids and Umayyads.” 
77.  Ibid., 56–7, 68, 70.
78.  At the convergence of the roads leading to Homs, Tiberias, Sidon, Tyre, and Beirut (Chehab, 

“Identification,” 43).
79.  None of the other Umayyad madīnas—let alone qaṣrs—of Greater Syria have a mosque‒dār al-imāra 

complex matching the one in ʿAnjar, namely a dār al-imāra built along the qibla of a sizeable congregational 
mosque and provided with two doors connected to two doors in the qiblī wall of the mosque, normally to the 
right and to the left of the miḥrāb. In most madīnas (and normally in qaṣrs), the mosques are embedded in 
substantial palatine cities (as in Qaṣr al-Ḥayr al-Sharqī, the citadel of ʿAmmān, or Mshatta); in others, such as 
ʿAqaba and Ramla, the case of a palace by the qibla side of the mosque has yet to be made by archaeology. This 
is particularly significant in the case of Ramla, which was expressly founded to serve as the capital of Filasṭīn, 
and the main residence of the caliph Sulaymān b. ʿAbd al-Malik. Recently, Gideon Avni reported the discovery, 
some eighty meters south-west of the mosque, of the remains of an opulent mansion which he tends to identify 
with Sulayman’s dār al-imāra. The presence of a later Islamic cemetery on the qibla side of the mosque, and 
the fact that the qiblī wall was rebuilt and partially re-oriented at a later date, however, prevented the scholar 
from identifying the existence of a “canonical” mosque‒dār al-imāra pairing there; see G. Avni, The Byzantine-
Islamic Transition in Palestine: An Archaeological Approach (New York: Oxford University Press, 2014), 165–7.

80.  D. Genequand, Les établissements des élites omeyyades en Palmyrène et au Proche-Orient (Beirut: 
Institut Français du Proche-Orient, 2012), 154–9, 242–8.



260  •  AilA SAnti

Al-ʿUṣūr al-Wusṭā 30 (2022)

residential functions, ʿAnjar is distinguished by the presence of a mosque‒dār al-imāra 
complex shaped upon a genuinely imperial model, such as is typically found in amṣar such 
as Kūfa, Baṣra, and Wāsiṭ; imperial capitals such as Damascus (and, later, Baghdad); or 
cities endowed with exceptional religious and political significance such as Medina81 and 
Jerusalem. This suggests that ʿAnjar was conceived as a reduced-scale, ideal model of an 
imperial capital, perhaps designed to host some important ceremony, or simply to evoke, 
symbolically, an imperial ritual setting.

This would add a further shade of significance to the phenomenon studied by Patricia 
Crone,82 according to which Umayyad ex-novo settlements in Bilād al-Shām were functional 
in gathering personal entourages of mawālī and soldiers who could support their lord in war 
during the violent succession conflicts that bedeviled the Marwānid state.83 Considered in 
the context of this functional interpretation, what could have been the meaning of a madīna 
whose symbolical core recreated the typical configuration of religious-administrative 
complexes of more substantial urban entities? Which kind of ceremonial required such 
configuration, and why ʿAnjar, of all the madīnas?

Regrettably, these and the other issues raised will remain unaddressed, and the 
hypothesis put forward unproven, until new archaeological evidence is made available. It 
is hoped that further investigations will be soon carried out at the site and the surrounding 
area, in the expectation that new light will be shed on the multifaceted function and rich 
history of ʿAnjar, which remain major enigmas for scholars in the field. 

81.  The presence of a mosque‒dār al-imāra pairing in Medina following the precise configuration described 
above (n. 79) seems to have existed in the city of the Prophet from at least the Marwānid period (Fig. 15). This 
is clearly emerging from an ongoing research project on the urban topography of Medina, which I am carrying 
out and the results of which I intend to publish in a forthcoming monograph. 

82.  P. Crone, Slaves on Horses: The Evolution of the Islamic Polity (Cambridge: Cambridge University Press, 
1980).

83.  A. Northedge, “Early Islamic Urbanism,” in Companion to Islamic Art and Architecture, First Edition, ed. 
F. B. Flood and G. Necipoğlu, 155–76 (Hoboken: Wiley Blackwell, 2017), 162.



Al-ʿUṣūr al-Wusṭā 30 (2022)

The Mosque‒Dār al-Imāra Complex at ʿAnjar •  261

Bibliography

Abou Diwan, G., and J. Doumit. “Ancient Wetlands of the Biqāʿ: A Buffer Zone between the 
Hinterlands of Sidon and Berytus in the Roman Period.” Bulletin d’archéologie et 
d’architecture libanaises 16 (2016): 215–52.

Almagro, A., and P. Jiménez. “The Umayyad Mosque of the Citadel of ʿAmmān.” Annual of the 
Department of Antiquities of Jordan 44 (2000): 459–75.

Arce, I. “Romans, Ghassanids and Umayyads and the Transformation of the Limes Arabicus: 
From Coercive and Deterrent Diplomacy Towards Religious Proselytism and Political 
Clientelarism.” In La Transgiordania nei secoli XII-XIII e le ‘frontiere’ del Mediterraneo 
medievale, edited by G. Vannini and M. Nucciotti, 55–74. Oxford: BAR Publishing, 2012.

Avni, G. The Byzantine-Islamic Transition in Palestine: An Archaeological Approach. New 
York: Oxford University Press, 2014.

Bacharach, J. L. “Marwanid Umayyad Building Activities: Speculations on Patronage.” 
Muqarnas 13 (1996): 27–44.

al-Balādhurī. The Origins of the Islamic State, Being a Translation from the Arabic, 
Accompanied with Annotations, Geographic and Historic Notes of the Kitâb Futûḥ 
al-Buldân of al-Imâm Abu-l ʻAbbâs, Aḥmad ibn-Jâbir al-Balâdhuri. Part 2, translated by 
F. C. Murgotten. PhD dissertation, Columbia University, 1924.

Ben-Dov, M. In the Shadow of the Temple: The Discovery of Ancient Jerusalem. New York: 
Harper & Row, 1985.

Bonatz, D., N. Ali, and C. Jauss. “Preliminary Remarks on an Archaeological Survey in the 
Anti-Lebanon.” Bulletin d’archéologie et d’architecture libanaises 16 (2002): 283–307.

Brooks, E. W. “A Syriac Chronicle of the Year 846.” Zeitschrift der Deutschen Morgenländischen 
Gesellschaft 51, no. 4 (1897): 569–88. 

Burdajewicz, M. “Some Remarks on the Architectural Decoration of the North-West Church 
at Hippos (Susita).” In Ex Oriente Lux: Studies in Honour of Jolanta Młynarczyk, edited 
by K. Jakubiak and A. Łajtar, 345–362. Warsaw: University of Warsaw Press, 2020.

Chehab, H. “On the Identification of ʿAnjar (ʿAyn al-Jarr) as an Umayyad Foundation.” 
Muqarnas 10 (1993): 42–8.

———. “Les palais omeyyades d’Anjar, résidences princières d’été.” Archéologia 87 (1975): 
18–25.

———. Les ruines d’Anjar, revu et annoté par G. Homsy-Gottwalles. London: Lebanese British 
Friends of the National Museum, 2017–18.

Chehab, M. “The Umayyad Palace at ʿAnjar.” Ars Orientalis 5 (1963): 17–25.



262  •  AilA SAnti

Al-ʿUṣūr al-Wusṭā 30 (2022)

Creswell, K. A. C. Early Muslim Architecture: Umayyads, A.D. 622–750. 2nd ed. Vol. 1, part 2. 
New York: Hacker Art Books, 1979.

Crone, P. Slaves on Horses: The Evolution of the Islamic Polity. Cambridge: Cambridge 
University Press, 1980.

Cytryn-Silverman, K. “The Umayyad Mosque of Tiberias.” Muqarnas 26 (2009): 37–61.

Di Cesare, M. “A Note on an Umayyad Carved Ivory Plaque Kept at the Walters Art Gallery.” 
Vicino Oriente 21 (2017): 197–210.

Duva, F. “Persistenza e sviluppo del dispositivo ‘moschea del venerdì–dār al-ʾimāra’ nell’area 
iranica in periodo proto-abbaside.” PhD dissertation, Sapienza Università di Roma, 2021.

Ettinghausen, R. From Byzantium to Sasanian Iran and the Islamic World: Three Modes of 
Artistic Influence. Brill: Leiden, 1972. 

Finster, B. “ʿAnjar: spätantik oder früislamisch?” In Residences, Castles, Settlements: 
Transformation Processes from Late Antiquity to Early Islam in Bilad al-Sham, 
Proceedings of the International Conference held at Damascus, 5-9 November 2006, 
edited by K. Bartl and A. R. Moaz, 229–42. Rahden: Verlag Marie Leidorf, 2008. 

———. “Researches in ʿAnjar, I. Preliminary Report on the Architecture of ʿAnjar.” Bulletin 
d’archéologie et d’architecture libanaises 7 (2003): 209–44.

———. “Researches in ʿAnjar II. Preliminary Report on the Ornaments of ʿAnjar.” Bulletin 
d’archéologie et d’architecture libanaises 11 (2007): 143–65. 

———. “Vine Ornament and Pomegranates as Palace Decoration in ʿ Anjar.” In The Iconography 
of Islamic Art, Studies in Honour of Robert Hillenbrand, edited by B. O’Kane, 143‒62. 
Edinburgh: Edinburgh University Press, 2005.

Fischer-Genz, B., and G. Ehrig. “First Results of the Archaeological Survey Project in the 
Territory of Ancient Heliopolis-Baalbek.” Bulletin d’archéologie et d’architecture 
libanaises 9 (2005): 135–38.

Flood, F. B. The Great Mosque of Damascus: Studies on the Makings of an Umayyad Visual 
Culture. Leiden: Brill, 2001.

Foote, R. M. “An Abbasid Residence at al-Humayma.” In Byzantium and Islam, Age of Transition, 
7th–9th Century, edited by H. C. Evans and B. Ratliff, 221–22. New York: Metropolitan 
Museum of Art, 2012.

Fowden, E. K. “Christian Monasteries and Umayyad Residences in Late Antique Syria.” 
Antigüedad y Cristianismo 21 (2004): 565–81.

Gordon, M. “Ibn Ṭūlūn, al-Qaṭāʾiʿ and the Legacy of Samarra.” Beiträge zur Islamischen Kunst 
und Archäologie 4 (2014): 63–77.



Al-ʿUṣūr al-Wusṭā 30 (2022)

The Mosque‒Dār al-Imāra Complex at ʿAnjar •  263

Grabar, O. “Ceremonial and Art at the Umayyad Court.” PhD dissertation, Princeton University, 
1955.

———. The Formation of Islamic Art: Revised and Enlarged Edition. New Haven, CT: Yale 
University Press, 1987.

———. “Umayyad Palaces Reconsidered.” Ars Orientalis 23 (1993): 93–108.

Genequand, D. “Les décors en stuc du bâtiment E à Qaṣr al-Ḥayr al-Sharqī.” Syria 88 (2011): 
351–78.

———. Les établissements des élites omeyyades en Palmyrène et au Proche-Orient. Beirut: 
Institut Français du Proche-Orient, 2012.

———. “Two Possible Caliphal Representations from Qaṣr al-Ḥayr al-Sharqı̄ and Their 
Implication for the History of the Site.” In Power, Patronage, and Memory in Early 
Islam: Perspectives on Umayyad Elites, edited by A. George and A. Marsham, 147–74. 
New York: Oxford University Press, 2018.

Hamilton, R. W. Khirbat al-Mafjar: An Arabian Mansion in the Jordan Valley. Oxford: Clarendon 
Press, 1959.

———. The Structural History of the Aqsa Mosque: A Record of Archaeological Gleanings from 
the Repairs of 1938–1942. London: Oxford University Press, 1949.

Heidemann, S. “The Evolving Representation of the Early Islamic Empire and Its Religion on 
Coin Imagery.” In The Qurʾān in Context: Historical and Literary Investigations into 
the Qurʾānic Milieu, edited by A. Neuwirth, N. Sinai, and M. Marx, 149–95. Leiden: Brill, 
2010.

Hillenbrand, R. “ʿAnjar and Early Islamic Urbanism.” In The Idea and Ideal of the Town 
between Late Antiquity and the Early Middle Ages, edited by G. P. Brogiolo and B. Ward-
Perkins, 58–98. Leiden: Brill, 1999. 

Jaia, A., and L. Ebanista. “Excavations at Istakhr in 2012: The Test West of the Site of the 
Mosque.” In Istakhr (Iran), 2011–2016. Historical and Archaeological Essays, edited by 
M. V. Fontana, 303–44. Rome: Sapienza Università di Roma, 2018.

Kennedy, H. The Prophet and the Age of the Caliphates: The Islamic Near East from the Sixth 
to the Eleventh Century. London: Routledge, 2015.

Krencker, D., and W. Zschietzschmann. Römisch Tempel in Syrien. Berlin: de Gruyter, 1938.

Labisi, G. Dwelling Models of Umayyad Madāʾin and Quṣūr in Greater Syria. Oxford: BAR 
Publishing, 2020.



264  •  AilA SAnti

Al-ʿUṣūr al-Wusṭā 30 (2022)

Leal, B. “ʿAnjar: An Umayyad Image of Urbanism and Its Afterlife.” In Encounters, Excavations 
and Argosies: Essays for Richard Hodges, edited by J. Moreland, J. Mitchell, and B. Leal, 
172–89. Oxford: Archaeopress, 2017.

Marfoe, L. Between Qadesh and Kumidi: A History of Frontier Settlement and Land Use in the 
Biqa, Lebanon. PhD dissertation, University of Chicago, 1978.

Marsham, A. Rituals of Islamic Monarchy: Accession and Succession in the First Muslim 
Empire. Edinburgh: Edinburgh University Press, 2009.

Massignon, L. “Explication du plan de Baṣra.” In Opera Minora, vol. 2 of 3, edited by Y. 
Moubarac, 61–87. Beirut: Dar Al Maaref, 1963.

al-Masʿūdī. Les prairies d’or. Edited and translated by C. Barbier de Maynard. Paris: Imprimerie 
Impériale, 1869.

Meritt, L. S. “The Geographical Distribution of Greek and Roman Ionic Bases.” Hesperia 38 
(1969): 186–204.

Miles, G. C. “ʿAnaza.” In The Encyclopaedia of Islam, 2nd ed., ed. P. Bearman et al., 1:482. 
Leiden: Brill, 1954–2009. 

———. “Miḥrāb and ʿAnazah: A Study in Islamic Iconography.” In Early Islamic Art and 
Architecture, edited by J. M. Bloom, 149–66. London: Routledge, 2002.

Molinari, A. Segesta II: Il Castello e la Moschea (scavi 1989-1995). Palermo: Flaccovio, 1997.

Munt, H. The Holy City of Medina, Sacred Space in Early Islamic Arabia. New York: Cambridge 
University Press, 2014.

Newson, P. “Archaeological Landscapes of the Bekaa: Past Results and Future Prospects.” 
Berytus 56 (2016): 257–80.

——— and R. Young. “The Archaeology of Conflict-Damaged Sites: Hosn Niha in the Biqaʿ 
Valley, Lebanon.” Antiquity 89 (2015): 449–63.

Nödelke, T. “Zur Topographie und Geschichte des Damascenischen Gebietes und der 
Haurângegend.” Zeitschrift der Deutschen Morgenländischen Gesellschaft 29/34 
(1876): 419–44.

Northedge, A. “Early Islamic Urbanism.” In Companion to Islamic Art and Architecture, First 
Edition, edited by F. B. Flood and G. Necipoğlu, 155–76. Hoboken: Wiley Blackwell, 2017. 

Pedersen, J. “Khaṭīb.” In The Encyclopaedia of Islam, 2nd ed., ed. P. Bearman et al., 4:1,109–
1,111. Leiden: Brill, 1954–2009.

Rey-Coquais, J. P. “Notes de géographie syrienne antique.” Mélanges de l’Université Saint-
Joseph 40 (1964): 287–312.



Al-ʿUṣūr al-Wusṭā 30 (2022)

The Mosque‒Dār al-Imāra Complex at ʿAnjar •  265

Rosen-Ayalon, M. “The White Mosque of Ramla: Retracing Its History.” Israel Exploration 
Journal 56 (2006): 67–83.

Sack, D. Resafa IV: Die Groẞe Moschee von Resafa. Mainz: P. von Zabern, 1996.

Safar, F. Wāsiṭ: The Sixth Season’s Excavations. Cairo: Imprimerie de l’Institut Français 
d’Archéologie Orientale, 1945.

Santi, A. “ʿAnjar in the Shadow of the Church? New Insights on an Umayyad Urban Experiment 
in the Biqāʿ Valley.” Levant 50, no. 2 (2019): 267–80. 

———. “Early Islamic Kūfa in Context: A Chronological Reinterpretation of the Palace, with 
a Note on the Development of the Monumental Language of the Early Muslim Élite.” 
Annali, Sezione Orientale 78 (2018): 69–103.

———. “Reinterpreting the Miṣr: New Insights for a Revised Chronological Attribution of the 
Mosque–Dār al-Imāra Combination in Kūfa.” In Proceedings of the 11th ICAANE, vol. 
2, edited by O. Adelheid, M. Herles, K. Kaniuth, L. Korn, and A. Heidenreich, 509–20. 
Wiesbaden: Harrassowitz Verlag, 2020.

———. “The Role of Madīna in the Emergence of the Mosque‒Dār al-Imāra Combination: A 
Preliminary Note.” Vicino Oriente 21 (2017): 211–23.

Sauvaget, J. La mosquée omeyyade de Médine. Paris: Institut Français de Damas, 1947.

———. “Notes de topographie omeyyade.” Syria 24 (1944–45): 96–112.

———. “Les ruines omeyyades de ʿAndjar.” Bulletin du Musée de Beyrouth 3 (1939): 5–11.

Sayyid, A. F. La capitale de l’Égypte jusqu’à l’époque fatimide, al-Qāhira et al-Fusṭāṭ. Stuttgart: 
Franz Steiner Verlag, 1998.

Scerrato, U. “Research on the Archaeology and History of Islamic Art in Pakistan: Excavation 
of the Ghaznavid Mosque on Mt Rāja Gīrā. Second Report, 1986.” East and West 36, no. 
4 (1986): 496–511.

Schlumberger, D. Qasr el-Heir el-Gharbi. Paris: P. Geuthner, 1986.

Schmitt, G. “Zum Königreich Chalkis.” Zeitschrift des Deutschen Palästina-Vereins 98 (1982): 
110–24.

Seyrig, H. “Les dieux syriens en habit militaire.” Annales archéologiques arabes syriennes 21 
(1971): 67–70.

Stillman, Y. K. Arab Dress: A Short History from the Dawn of Islam to Modern Times. Edited 
by N. A. Stillman. 2nd ed.  Leiden: Brill 2003.

———. “Libās.” In The Encyclopaedia of Islam, 2nd ed., ed. P. Bearman et al., 5:732–42. Leiden: 
Brill, 1954–2009.



266  •  AilA SAnti

Al-ʿUṣūr al-Wusṭā 30 (2022)

Swelim, T. Ibn Tulun: His Lost City and Great Mosque. Cairo: AUC Press, 2015. 

al-Ṭabarī. The History of al-Ṭabarī (Taʾrīkh al-rusul waʾl mulūk). Vol. 24, The Empire in 
Transition, translated by D. S. Powers. Albany, NY: State University of New York Press, 
1989.

Theophanes. Theophanis Chronographia. 2 vols. Bonn: Weber, 1839–41.

Theophrastus. Enquiry into Plants and Minor Works. Vol. 2, On Odours. Weather Signs, 
translated by A. Hort. London: W. Heinemann; New York: G. P. Putnam’s Sons, 1916.

Treadwell, L. ““Mihrab and ʿAnaza” or “Sacrum and Spear”? A Reconsideration of an Early 
Marwanid Silver Drachm.” Muqarnas 22 (2005): 1–28.

Walmsley, A. The Danish-Jordanian Islamic Jarash Project, Season Report 2008. Copenhagen: 
Copenhagen University Press, 2008. 

———. “The Friday Mosque of Early Islamic Jarash in Jordan.” Journal of the David Collection 
1 (2003): 111–31.

——— and K. Damgaard. “The Umayyad Congregational Mosque of Jarash in Jordan and its 
Relationship to Early Mosques.” Antiquity 79 (2005): 362–78.

Whitcomb, D. “The City of Istakhr and the Marvdasht Plain.” In Akten des VII. Internationalen 
Kongresses für Iranische Kunst und Archäologie: München, 7.-10. September 1976, 
363–70. Berlin: D. Reimer, 1979.

Wilkinson, C. K. Nishapur: Some Early Islamic Buildings and their Decoration. New York: The 
Metropolitan Museum of Art, 1986.