MACHIEL KIEL

THE QUATREFOIL PLAN IN OTTOMAN ARCHITECTURERECONSIDERED IN LIGHT OF THE "FETHIYE MOSQUE" OF ATHENS

This type, almost certainly a creation of the Ottomancapital is best expressed in the great mosques of Istanbul,but it also occurs in Algiers, throughout the Balkans,in Syria, and even in Muhammad Ali's Cairo. It doesnot occur in Morocco, Iran, India or Central Asia be-cause this type is tied to Ottoman supremacy. It servesan Islamic function, but its architectural forms signifya specific empire.

Oleg Grabar, "The Iconography of Islamic Architecture."

Ottoman architecture celebrated some of its greatesttriumphs in three prominently sited Istanbulmosques-the Sehzade Mehmed mosque (1543-48),the Sultan Ahmed I or Blue mosque (1609-17), andthe Yeni Cami (begun in the 1590's, completed in1666)-built according to the quatrefoil or clover-leafcross-in-square plan. This plan type gained great popu-larity, and representative examples of it can be seenfrom Diyarbaklr in southeastern Anatolia to the cityof Tunis and the island of Djerba at the extreme west-ern boundary of the Ottoman world, with at least threeexamples in Greece as an intermediary stage. Whenin 1833 Muhammad 'Ali started work on his enormous"Alabaster Mosque" on top of the citadel hill in Cairohe also chose the quatrefoil plan. When in the 1970'sit was decided that Republican Ankara was to have an"imperial mosque" the quatrefoil plan was once againchosen, showing its unbroken appeal throughout theages.

Over more than half a century a number of schol-ars have put forward theories about the plan's origin,but the definitive answer had still to be found. Thisessay reconsiders the elusive and controversial sourcesof the plan type along with later versions, after con-sidering some of the arguments about its origin in thesecondary literature. If we disregard the somewhatcrude formulation of the Dutch Orientalist H. J.Kramers, who declared the ehzade mosque to be "twoHagia Sophias interlocked and fused together," a va-riety of possible hypotheses remain. In reviewing them

we shall see how they influenced each other, what theyleft out, and where their arguments go astray. Finallywe will suggest what we think are the true origins ofthis remarkable plan.

In 1953, the Swiss-trained Turkish scholar UlyaVogt-G6knil saw the plan of the ehzade mosque first ofall as resulting from a confrontation with the HagiaSophia and noted:

With him [Sinan], ... Turkish architecture reaches thesummit of its development. Around 1548, forty years afterthe mosque of Beyezid II, he again went back to theHagia Sophia for his first really great building, the

ehzade mosque.... Within fifteen years Sinan createdthree completely new variants of Hagia Sophia's plan-scheme. In the ehzade mosque the bilateral symme-try of Hagia Sophia was transformed into a radially sym-metrical one, in other words a quatrefoil plan.'

Six years later the Turkish art historian Behqet Unsalfound a very different origin for the plan of the Sehzademosque, suggesting, though not in so many words, thatthe great Byzantine church had nothing to do withit:

[The plan] of the ehzade Mosque, on the other hand-a central dome surrounded by four half-domes-has itsforerunner, not in Istanbul, but at Maras in the coun-try of the Dulkadir family, namely the Ulucami atElbistan repaired by Alafiddevle in 1479-1515. Turkisharchitects worked for a century and a half on this typeof building. The master himself, when he began hisdesign for the ehzade Mosque, aimed at giving monu-mental form on the lines of the old Turkish buildingshe had seen in his youth and, while introducing inno-vations to the capital, yet maintaining ties with tradition. 2

By the "century and a half" of building experienceOnsal evidently meant experimenting with the cen-tral domed mosque to which a half-dome, housing themihrab, had been added, as in Mehmed II's old Fatihmosque (1463-70) in Istanbul and the Yahsl Bey zwiya(convent) mosque (1441) in Tire in western Anatolia,

� _I I

MACHIEL KIEL

or the Beylerbey zdwiya mosque (1428-29) in Edirne. 3

Although Unsal pointed to the Elbistan connection,he did not develop it further. He also seems to haveforgotten that the young Sinan could not have seenElbistan. While marching with the army in the cam- paign of the two Iraqs ('Iraqayn) in 1534-35, he didnot pass through Elbistan at all. The army marchedmuch further to the north, via Konya, Kayseri, Sivas,Erzincan, Erzurum to Tabriz and Hamadan in Iranand from there to Baghdad, and after a second cam-paign in Iran in 1535 returned by way of Diyarbakir,Urfa, Halep and Adana.4 The closest Sinan came toElbistan was five full days' journey away.

In 1970, Metin Sozen took a great step forward withthe publication of an article which introduced amosque built using a quatrefoil plan, but, in contrastto Elbistan, clearly dated. It was the mosque of SinanPasha in the small town of Hacl Hamza in northernAnatolia, a way station on the road from Tosya toOsmanclk and Merzifon. This mosque was built in1506-7 and then unfortunately ruined in an earth-quake and more or less repaired in a different form.After S6zen had published his remarks, it was replacedby a wholly new building.

In the same study S6zen pointed to the existenceof a similar building in the vicinity of Hacl Hamza. 5

This is the mosque of Oguzkty in the district (ndhiye)of Kargi, which was apparently built a decade or twolater than the Hacl Hamza mosque. S6zen only report-ed the existence of the Ouzk6y building; he did notprovide details or a plan. I studied the building in detailin 1998. According to local tradition, it was a founda-tion of the sixteenth-century vizier Goban MustafaPasha (d. 1529) known for his huge complexes (killiye)in Gebze, between Istanbul and Izmit, and in Eskisehir.His buildings in Oguzkty originally included a khanand a hammam and must have been intended as ahalting station on the Kastamonu-Tosya-Merzifonhighway. The khan and hammam were swept awayduring a flood some thirty years ago, but they werewell remembered by the local population. The stabil-ity of the mosque was seriously affected by the sameflood and a landslide that destroyed the other build-ings. To keep it standing, it needed the help of enor-mous buttresses on three sides.

The buildings of Coban Mustafa Pasha in Oguzktyare not mentioned in the waqf section of the 1530census (tahrir) of northern Anatolia. However, as thebasic data for this "proto-statistical" survey of the wholeempire were collected in 1520-21, in the very first years

of the reign of Sfileyman the Magnificent, the build-ings must be later.6 The vizier's little complex inOguzk6y, therefore, must be dated between 1522-23and 1528-29. The mosque shows a truncated versionof the quatrefoil plan. It has a central dome and threehalf-domes resting on two, instead of four, piers. Themosque is built of carefully executed cloisonnee, butas a whole it conveys a rustic impression. It is certainlythe work of a provincial master. For architectural his-tory it is of importance only as an early example ofthe plan, which,judging by the number of examples,must have enjoyed considerable popularity in thesenorth Anatolian districts.

A year after S6zen, Godfrey Goodwin first publishedhis classic Histo7y of Ottoman Architecture. Goodwinpicked up where Unsal left off, at the mosque of Elbi-stan, which he studied himself. Goodwin concludedrightly that Unsal had evidently never actually beenin Elbistan. 7 He noted: "At some time in the reign ofBayezid II, between 1479 and 1515 the Ulu Cami wasrebuilt at Elbistan in the territory of the Dulkadir fam-ily, southhwest of Malatya. .... This mosque is evi-dence- but in no way as conclusive as Unsal asserts-of the emergence of the quatrefoil plan in Anatoliaprior to Sinan's building of ehzade. Indeed the qua-trefoil plan existed in Central Asia long before."8 Thelatter thought is not further worked out, but thenGoodwin introduces a very interesting link, pointingto the works of the first grand vizier of Sultan Sfileyman,Piri Mehmed Pasha, who built a number of interest-ing mosques and masjids:

At Hask6y on the Golden Horn, he built his majormosque, which is now a delapidated tobacco warehouse;it was heavily restored in the nineteenth century. It wasbuilt on a grand scale (in 930/1523, the year of his fall)and was the first quatrefoil mosque in Istanbul. Thiscentralized, four-leaf-clover plan was the accepted cli-max and perfection of the ideal Ottoman mosques whichall the previous architects had been led towards andwhich only Sinan was to eclipse. It was ante-dated by themosque of Fatih Pasha at Diyarbaklr, built between 1518and 1520, perhaps in the winter when the army was idlein its quarters.... To Diyarbakir came an unknownarchitect who, prior to his patron's death in 1521, wasto achieve the Bflyfik Mehmet or Fatih Pasha, the firstcentralized Ottoman mosque, with four semi-domes andnot mere vaults.9

Goodwin evidently did not know Metin S6zen's articleand therefore did not know of the Hacl Hamza, whichwas fifteen years older. In his description of the ehzade

- ------

110

THE QUATREFOII, PLAN IN OTTOMAN ARCHITECTURE: THE "FETHIYE MOSQUE" OF ATHENS

Fig. 1. Oguzk6y (Kargi/Tosya district). Mosque of oban Mustafa Pasha, 1520's. Photo: M. Kiel 1998.

mosque, Goodwin totally rejected the influence of theHagia Sophia, connecting it instead, via the Piri Pashamosque in Hask6y, with Diyarbaklr:

In November 1534 his sultan commanded the buildingof a great complex at Sehzadebasl in honour of theprince [Mehmed] .... Sinan made no attempt to excelHagia Sophia but was absorbed by the concept of thecentralized dome and turned to a plan like that of FatihPasha at Diyarbakir or Piri Pasha at Hask6y, and theancient tradition of which these mosques' form was thenthe climax. 0

In 1981, ten years after Goodwin's book was published,Sauermost and Von der Mfilbe published their monu-mental work on the mosques of Istanbul. In theirdescription of the behzade mosque they point to par-allels with contemporary Italian Renaissance architec-ture but exclude direct influence. Then they offer amuch broader explanation of the origin of the plan,

linking it to older concepts like that of Hagia Sophia,or an Armenian background, with Unsal's, S6zen's andGoodwin's ideas:

Early medieval quatrefoil buildings in Armenia couldhave inspired the idea of ordering the vaulting systemof the Hagia Sophia in a cross-shape because the old-est preserved mosque showing this vaulting concept stillstands in Diyarbaklr in eastern Anatolia, then a garri-son near Armenia. The thought could also have comevia the many Byzantine cross-in-square churches. In1413-21 Sultan Mehmed I erected a mosque in the townof Dimetoka, 40 km south of Edirne, following this plan,albeit without half-domes.... The Fatih Pa§a Camii inDiyarbakir (1518-22) departs from the vaulting systemof the Hagia Sophia and also brings in the four radialhalf-domes. The inner space, however, remains old fash-ioned and primitive. After one or two successor build-ings the new trend ended for a time. Then in 1523 GrandVizier Piri Mehmet Pasa brought the type to Istanbul.

�1� �

111

MACHIEL KIEL

Sinan must have known his now profaned mosque inHaskty on the northern shore of the Golden Horn.

It has to be added that the Dimetoka mosque of Meh-med I (1413-21) does not fit into this otherwise en-lightening synthesis. In the sixteenth-century tahrfrdefters this mosque appears as a foundation of YlldrsmBayezid (r. 1389-1402). Today it appears as a cheaplyexecuted quatrefoil plan, with four massive piers car-rying no domes and half-domes, but crudely executeddome-like vaults of wood. Ayverdi, after careful studyof the building by one of his assistants, concluded thatthe original plan must have been very different: twopiers, supporting two big domes over a central naveand two cradle vaults over each of the lateral naves.12

In 1668, the Ottoman traveler Evliya Qelebi had re-marked that the covering of this mosque was madeof wood. Recently the dendrochronological work ofPeter Kuniholm and Lee Striker has established with-out any doubt that the oak used in the upper part ofthe building was cut in 1419. The enormous beamscarrying the vaults date from 1439. The great inscrip-tion over the central entrance of the mosque gives theyear 823 (1420); the one over the lateral entrance says824 (1421). The conclusion is simple: Bayezid's mosqueremained without vaulting at his death in 1402-3 andhis son Mehmed completed it with a rather smallbudget, but had his name written on the inscriptions,and ignored the work of his predecessor. In the samemanner he had completed the famous Eski Cami innearby Edirne, which was begun by Mehmed's brother,Amir Sfileyman, and continued by his successor MusaQelebi; it was completed in one year by Mehmed I,who is the only ruler whose name appears on the in-scriptions.l 3 The wood of 1439 is connected with amajor repair. The Dimetoka mosque is thus not anearly example of the quatrefoil plan.

In his Osmanlz Devri Mimarisi (1986), the veteranTurkish art historian Oktay Aslanapa picks up whereUnsal left off. He emphasizes the interest Turkisharchitects showed in the half-dome from the time ofSultan Murad II and through the buildings constructedduring the reigns of his sons-Mehmed II followedby Bayezid II-in the new centers of Turkish archi-tecture, Diyarbakr and Elbistan, until under Selim Ithe centralized plan with the four half-domes was fi-nally conceived. It should perhaps be remembered thatthe ruling house of Elbistan was the Tfirkmen dynastyof Dulkadir (Dhu'l-Kadr) and that Diyarbakir had beenone of the most important centers of the equally

Turkish dynasty of the Akkoyunlu. Aslanapa correctlyremarks that the inscription of 637 (1239) on the Elbis-tan mosque, mentioning Amir Mfibarizuddin Cavli andthe Rum Seljuk ruler Giyasiiddin Keyhfisrev, was takenfrom somewhere else and placed in the masonry ofthe Dulkadirid mosque.' 4

In his description of the ehzade mosque, Aslanapapoints to the direct influence of buildings in Diyarbaklrand Elbistan:

Taking the problem of the half-dome in hand for thefirst time and setting aside the plan of the Hagia Sophiaand the Bayezid Mosque, Mimar Sinan achieved in thiswork [the ehzade], the ideal of a four-half-domedcentrally planned building, thereby realizing the dreamof the architects of the Renaissance. Upon seeing theUlu Cami of Elbistan and the Fatih Pasha Mosque inDiyarbaklr Sinan must have recognized the potential ofthe four-half-dome plan as incorporated in these twobuildings and used it to create something magnificent.' 5

Aslanapa, like Unsal, did not realize that Sinan couldnot himself have been in Elbistan.

In 1987, a year after Aslanapa's great survey, AptullahKuran published a major work on Sinan, in whichinescapably he also deals with the emergence of thequatrefoil plan.)6 Kuran skips over Elbistan withoutcomment but summarizes the role Diyarbakir mighthave played as follows:

I consider the Uskfidar Mihrimah Sultan a direct descen-dant of the Old Fatih [Mehmed II] Mosque. Likewise,the ehzade Mehmed has a similar kinship with theBayezid [II]. In these relationships both Sinan mosquesemulate the older ones with one significant difference:their two-domed side units flanking the central domedspace are surmounted by half-domes. In the design ofthe ehzade Mehmed, one other mosque could haveplayed a significant role. This is the Fatih Papa in Diyar-bakir which Sinan must have seen and studied duringthe Two Iraks Campaign when the Ottoman army restedfor three weeks in Diyarbakir. Built by Beylerbeyi BlylklhMehmed Pasa between 1516 and 1520, the Fatih Paa'scentrally domed superstructure with a half-dome on eachside and a small dome at each corner could have pro-vided the inspiration for the ehzade Mehmed.' 7

Forty years after her first study, Vogt-Gknil, in a majornew work, again returned to the ehzade: "The com-bination of a dome and three or four half-domes wasnot new. It had already been realized in the fourthand the fifth century in monastery churches in Sohag,Upper Egypt, and in the church of San Lorenzo in

_D�ll� � �_

112

THE QUATREFOIL PLAN IN OTTOMAN ARCHITECTURE: THE "FETHIYE MOSQUE" OF ATHENS

Milan. The same combination was later used in thegreat Romanesque churches of Cologne (St. Maria imKapitol, Gross St. Martin, and others). But in none ofthese buildings do the four piers, carrying the dome,stand free." 18 She then points to "some mosques fromthe first half of the 16th century, in Cankin, Elbistan,and Diyarbakir," and suggests that Sinan might haveseen them during the campaigns to Iraq and Iran.'9

Elbistan has to be ruled out, as mentioned earlier. TheGankiri mosque is from 1558 and thus also has to beruled out, but as a whole Vogt-Gbknil's comparativeframework greatly enriches our understanding of theproblem. It should be added, however, that in thegrand church of Maria im Kapitol in Cologne the maincentral dome is carried by four free-standing piers, butthis, of course, does not make this church from theyear 1030 a remote ancestor of our mosque type.

In a fine new synthesis of Sinan's work, anotherveteran Turkish art historian, Dogan Kuban, proposeda wholly different solution to the problem of the ori-gin of the plan. Accusing foreign historians of art ofcreating a "myth" based on "purely superficial obser-vation," he boldly states:

In large domed structures, a fully symmetrical supportsystem is always an ideal. In the course of world archi-tecture, numerous domed buildings have been con-structed with centralized plans. Ottoman architects, whoconsistently used the dome as the covering element, werebound to employ this scheme at some point. In $ehzadeSinan gave this symmetrical scheme, also used by Re-naissance architects, a shape in conformity with thetraditions of Ottoman architectures

This is a valid line of argument, but in this particularcase does not explain everything. As Grabar remarkedin the quotation that begins this article, the plan wasnot used in Iran or Mughal India or in the Maghrib(or in Mamluk Syria and Egypt, one might add), al-though the architecture of these countries made fre-quent use of domes for covering spaces. If Kuban'sview is correct, then the Safavids, Mughals, and Mam-luks would also have arrived at the quatrefoil plan. Infact they did not.

We can now see a line of development more clearlythan before. Elbistari has to be skipped as a directsource for Sinan's work, but in one way or anothermay have played a more general role in the back-ground. The direct line seems to run from Diyarbaklrto 5ehzade. The existence of a mosque using the sameplan in 1506 in Hacl Hamza near Tosya, which is not

taken into account by anyone except Sozen, indicates,however, that the plan is older. It supports those whowould have Elbistan play an important (if indirect)role. The Elbistan building remains very problematicbecause no secure date can be found for it. In his article"Elbistan," Mehmet Tasdemir noted that the yearsbetween 1490 and 1505 were a flourishing period forElbistan. Alafiddevle, the ruler who according to Unsalhad the mosque "repaired," is reported to have builttwo mosques and one madrasa in Elbistan. In 1505,the Safavids took the town and destroyed it, where-upon the capital of the Dulkadir principality (beylik)was moved to Mara. Under Ali Bey (1515-22), thesuccessor of Alaiiddevle, it was moved to Elbistan,where Ali Bey inaugurated new building activity.21 Isthe mosque in its present form the product of a re-pair by Ali Bey? In his carefully written monographon the Dulkadir beylik, Refet Ymnanc (who is himselffrom Elbistan) made clear that it is not at all certainwhich of the old mosques of Elbistan is the one re-built (or reconstructed) by Alafiddevle, who indeedappears to have been an enthusiastic builder.2 2 Thusthe Elbistan mosque, although important, is not a goodstarting point for studying the great buildings inIstanbul.

The Abdurrahman Pasha mosque in Tosya in north-ern Anatolia, built, according to its inscription, in 992(1584), brings us back to the area of Hacl Hamza,where the oldest firmly dated example of the quatre-foil plan stood. The plan must have been particularlypopular in that area. Another example is the Ulu Camiof Cankiri, built, according its inscription, in 1557-58.23 Nowhere in Anatolia or the Balkans is there sucha concentration of examples of the plan as in north-ern and central Anatolia, the ancient province ofPaphlagonia. This is no coincidence: in the same area,and in the adjacent districts of Amasya, Merzifon,Tokat, and further south, there is a long tradition ofcovering a square space with a central dome restingon four free-standing supports. The way the supportsare placed is the same as in the quatrefoil mosques.All buildings in this tradition are of wood and all havebeen very little studied. It was particularly favored forthe congregational halls (cem evi) of the Alevis, a Shi'itesect prominent in central and northern Anatolia, whichwere built half underground to escape the punitiveeye of the Sunni Muslims. Since the 1980's, when theAlevis emerged into the open, the old cem evis wererapidly replaced by new constructions. Gfinkut Akinhas studied some of them and published a particularly

I� _ �

113

MACHIEI. KIEI.

Fig. 2. Tosya. Abdurrahman Pasha's Mosque, 1584. The now vanished mosque of Thebes, Greece, 1667/68 closely followedthis plan and building type. Photo: M. Kiel 1998.

fine example from Yahyall near Corum. Nezih Basgelenhas published some cruder examples north of Malatyaand mentioned having come across dozens of themin the villages of the same district, none of them stud-ied.2 4

Until recently, when it was replaced by a new struc--ture in much the same style, the main hall of theselamlzk, which was a hall of this kind, stood on MountAlnus, northeast of Tokat. It was the central place ofworship of the Hubyar Alevis and is many centuriesold. Hldlr Temel, the grandson of the last HubyarDede, now living in Cologne, possesses a large collec-tion of Ottoman sultanic orders, hiiccets, and other statepapers recording the property of the tekke (dervishconvent) through the ages. The oldest is from the earlyyears of Sultan Sfileyman's reign (1520-66), the new-est from the first years of the Republic. The villagemosque of the Hubyar center also shows the same plan

and construction. Characteristic is the great woodendome in the center of the building. It is a sophisti-cated structure called a krlang7,u kubbe in Turkish,Lanternedach in German, and "lantern roof' in English.This type of roofing is very old and widespread. It wasknown in China of the Han dynasty (206 B.C.-A.D. 220).A good example of it from the Ming period, datingfrom 1444, was taken from the Temple of WisdomChih-hua ssu in Beijing and can now be seen in thePhiladelphia Museum of Art. The Chinese term forthe vault is "Well of Heaven Ceiling" (G'ien ching), awell from which heavenly spirits and brightness comedown to those assembled below. This symbolism wouldfit very well in the Alevi context, but the connectionis, of course, difficult to prove. The kzrlan~gz kubbe iswell known in northern Pakistan and used for villagemosques as well as for larger rooms in houses. It canbe found in Caucasia, in Georgia, and Azerbaijan. In

_�lp�� �1�C� _

114

THE QUATREFOIL PLAN IN OTTOMAN ARCHITECTURE: THE "FETHIYE MOSQUE" OF ATHENS

the Sivas province, in Bing6l, and Mu§ the kzrlang-tkubbe on four free-standing supports is reportedly wide-spread.2 5 In the old houses of Erzurum it was used tocover the kitchen, the most important room in thehouse.2 6 All this implies that, in addition to a Byzan-tine or Armenian source behind the Ottoman quatre-foil plan suggested in the literature, one might considerold Anatolian and Asian building traditions, with thewooden kzrlangz¢ kubbe on four free-standing supportsas the element providing the initial spark. The un-known architects of Haci Hamza, Elbistan, and Diyar-baklr (or earlier works which have not come down tous) then translated the idea from wood into stone, awell-known development. Via Diyarbaklr it reachedSinan, who then brought it to its ultimate perfection.

One other mosque built according to the quatrefoilplan, which is mentioned rather frequently in the lit-erature, is regarded as possibly its oldest representa-tive. This is a mosque in the center of old Athens, whichbears the name Fethiye (Conquest) mosque. Athenswas added to the Ottoman dominions in 1456, andin 1458 Sultan Mehmed II himself visited it for severaldays, as we are told by his Greek panegyrist Kritoboulosand by the Athenean chronicler Chalkokondylas, bothcontemporaries to the event. For more than a gen-eration the Athens mosque has been mentioned in theteaching of Ottoman architecture at Istanbul Univer-sity as the oldest of the group. If true the building isof utmost importance, and sets us on a track very dif-ferent from eastern Anatolia.

The building was first introduced to the scholarlyworld by two well-known Greek art historians of thepast generation, Anastasis Orlandos and AndreasXyngopoulos,2 7 and made known in Turkey by SemaviEyice.28 In 1960, in the first of his two great mono-graphs on the monuments and urban developmentof Athens, which were later to become the standardworks on the subject, Ioannis Travlos stated that theFethiye mosque of Athens was a foundation of theconquering sultan and its construction had to be placedimmediately after the conquest in 1456. In his laterwork he stuck to this opinion.2 9 Thirteen years later,in 1973, the mosque was studied in detail by the in-defatigable Ayverdi, who also offered a detailed planand section. He called the mosque "Fatih Camii" andincluded it without much ado into his work on thearchitecture of the time of the Conqueror.30 Aslanapaalso incorporated it in his 1986 book as work fromthe time of Mehmed II,31 as did Nusret Qam in 2000,

in his luxuriously produced work on Ottoman archi-tecture outside Turkey.3 2

In 1991 Eyice came back to the Fethiye mosque inAthens, again remarking that the name of the build-ing suggested that it was from the time of the con-quest of Athens and its patron was Fatih SultanMehmed. It could, however, also date from the recon-quest of Athens, after the Venetian occupying force(September 1687-March 1688) had been driven away.Both Eyice and Ayverdi noted the very late and un-Ottoman features of the mosque, which would pointto the seventeenth or even the eighteenth century,rather than the fifteenth. Aslanapa declared that thefive-domed portico with round arches was added later.But Eyice concluded that "if the mosque is indeed fromthe time of the Conqueror, that is, from the mid fif-teenth century, then it is a very important buildingfor the history of Turkish architecture. 3 3 Importantit indeed is, but it is not from the fifteenth centuryand therefore not the oldest one known, but simply aseventeenth-century trend-follower. The answer to thequestion of its date is found in the sixteenth-centuryta.hrr defters of the sanjak of Egriboz, a valuable cat-egory of source material that is hopelessly neglectedby most architectural historians.

The oldest preserved tahnr of the sanjak of Egribozcontaining Athens is from 1506. According to thissource, the town of Athens then had (disregarding thegarrison of 75 men in the Acropolis castle) 1,716 Chris-tian and only one Muslim households. According tothe next tahrzr, in 1521 it had 2,286 Christian house-holds and 11 Muslim households. These numbers al-low us to conclude that in the lower town of Athensin the first three-quarters of the century there was noneed at all for a relatively large and sophisticatedmosque. The only mosque mentioned in the recordsis the famous Parthenon, at that time a Friday mosqueserving the needs of the garrison in the Acropolis cita-del.

The 1540 register is the first to mention Muslimbuildings in the lower town. In addition to the Cita-del mosque of Athens (cami'-i 'kal'e-i Atina), it men-tions the masjid of Yunus Voyvode and an elementaryschool (mu'allimhane) of Mehmed Voyvode. The 1570register shows that Islam was making inroads amongthe Athenians-by then the town had 57 Muslim house-holds. The waqf section of the same register shows thatthe increase in the number of Islamic buildings ranparallel to an increase in the number of Muslim in-habitants. First comes the only mosque of the town,

.- �----Oi�rprr�)�*1Oc� --- -- 1. -I-------------

115

MACHIEL KIEL

Elbistan, Ulu Cami, 1490-1505, afterS6zen

Diyarbakir, Fatih Pasa Camii, 1518-20, after S6zen

i Io 5 1ram.

Tosya, Abdurrahman PaSa Camii, 1584. M.K.2001

Oguzk6y (Kargl) Mosque of obanMustafa Pasa, 1522-28. M.K 2001

, ..-- I

Athens. "Fethiye" Mosque, 1668-70

Fig. 3. (All on same scale)

I�� _ L_ I _ _�

116

THE QUATREFOIL PLAN IN OTTOMAN ARCHITECTURE: THE "FETHIYE MOSQUE" OF ATHENS

the Parthenon/mosque of Sultan Mehmed Fatih onthe Acropolis, which had no waqf of its own. Its hattb,imam, and muezzin were paid from the jizya taxes ofthe district of Athens, the usual arrangement for smallsultanic foundations in the provinces. It is followedby the masjid of Yunus Voyvode in the lower town, themasjid and school of Memi Celebi ibn Tur Ali Aga anda school (mu'allimhane) of Mehmed Aga. A numberof very small waqfs provided for extra Qur'an recitalin the Acropolis mosque as well as for support of thesmall waqfs in the lower town.3 4 There is no sign what-soever of a Fethiye mosque.

This should be enough to declare the Fethiye Camiiof Fatih Sultan Mehmed in the lower town of Athensa "ghost mosque," but there is still another reason tovanquish forever the myth that this mosque is the oldestexample of the quatrefoil plan. The popular wisdomin Athens had it that the Fethiye mosque had previ-ously been the church of the Panaghia tou Staropa-zarou, or Church of the All Holy (Virgin Mary) of theWheat Market-stari here being the local pronuncia-tion of sitari (wheat), and not staro ("old" in Slavic)-and it was said to be a Byzantine building. As early as1929, George Sotiriou and Anastasis Orlandos showedthe theory to be groundless: the building was notByzantine but definitely Ottoman.3 5

Evliya Qelebi, who was in Athens in 1667, mentionsseveral mosques in the lower town of Athens by nameand enthusiastically describes the perfect beauty of theold Greek temple on the Acropolis, now a mosque.About a Fethiye, or Fatih, mosque in the lower townhe is silent.3 6 There is, however, an anonymous Otto-man travel account preserved in a mecmft'a in theOriental collections of the Bibliotheque Nationale inParis, which hitherto seems to have been overlooked.3 7

The style of writing of this text is late eighteenth cen-tury. The fact that it describes the citadel of Kara Babaoverlooking the town of Egriboz (the city of Chalkis,36 km east of Thebes) which was built immediatelyafter the unsuccessful siege of Venice in the autumnof 1687, places it definitely after 1687 (and before 1821,when the Ottoman period came to an end there). Anote toward the end of the text gives "ibu sene-i 1221"(1797). This text, written by a dervish traveler, nar-rates a story which had taken place long ago to ex-plain how a Christian church, as the result of a conflictbetween the two religious communities in Athens, wasconverted to a mosque. A qadi had been responsiblefor the action, and therefore the building becameknown as "'.kadi cami'i." The local Athenean Christians

had complained to the Porte in Istanbul but in theensuing legal struggle the qadi won the case. Thebuilding in question stood, according to our text, inthe Corn Market ("cdmi'-i mezkitr kapzsz niinde haftadabir def a bugday pazan kurulurfuktrtd dad u sitad eylerler").The qadi in question then added a mihrab, minbar,and minaret and provided for a waqf. This church canonly be the Panaghia tou Staropazarou, on the site ofwhich the Fethiye mosque now stands.

In 1676, Dr. Jacob Spon from Lyons and GeorgeWheler, later rector of Durham University, visitedAthens. After mentioning that the Christians of Ath-ens had two hundred churches in and around the town,of which fifty were used regularly, they noted: "TheTurks have five Mosques here, four in the Town andone in the Castle. The Mosque of the Bazar, in themiddle of the town, is the best of them." Later theyremark:

From the Temple of Augustus.. . you will come to theFront of a building over the Street, in an Entrance fash-ioned like a temple.... Passing through this templeEastwards you come into the Bazar, or Market-place;where on the right hand is a mosque, which, they say,was formerly the Cathedral Church. But it was rebuiltby the Turks since, and is altogether now a new fabric.3 8

This description is specific and detailed enough toexclude all other explanations. Furthermore, the name"Fethiye mosque" is definitely an old name and notan invention of historiography. It is mentioned in 1722,in the waqfiyya for the madrasa of Ruzname-i EvvelOsman Efendi, the ruins of which still stand oppositethe Tower of the Winds.39 If we now inspect the Fethiyemosque carefully we can see that the marble frame ofthe entrance portal and the marble frames of thewindows in the porch are covered with half-fadedOttoman inscriptions, pious wishes, and lines of po-etry. A number of them are dated. The oldest is from1080 (1669-70). Others follow closely in time. Theystrongly suggest that the name of the building is as-sociated with the final conquest of Crete by the Otto-mans.

In the 1797 story of our anonymous dervish theconfiscation of the church is also called a "conquest."The inter-communual tension that is echoed in theaccount of our anonymous traveler must be associatedwith the long Cretan War, which placed a severe strainon all levels of society. The old church was thus con-fiscated in the 1660's, knocked down soon after, andreplaced with the structure we see there now. This

___l_�_ll_·_PIII___C____�

117

MACHIEL KIEL

would immediately explain the pronounced un-Otto-man features in the architectural details-lotus capi-tals, round arches, round-arched windows, weakprofiles-which postdate the period of classical Otto-man architecture, when the state was slowly losing itsgrip on the outlying provinces and local architecturalpractices were growing correspondingly stronger. Thenew mosque was called "Fethiye" to celebrate the vic-tory in Crete in the same way as the Pammacharistoschurch in Istanbul was renamed the "Fethiye" to cel-ebrate the 1590 Ottoman victory in the Caucasus.

After World War II, archaeological excavations wereundertaken around the Athens Fethiye mosque. Theywere still going on in 2001, but remain unpublished.What we can see today are the foundations of the east-ern end of a church, with a half-round central apse,flanked by two small apses, of which the southern lat-eral apse still remains buried under the earth or ispartly under the mosque. This plan is typical for amiddle-Byzantine church. The mosque stands overthese ruins at an angle because it is oriented towardsMecca (here east-south-east, at 130 degrees of thecompass), whereas churches are always oriented dueeast (90 degrees), as is required by Greek Orthodoxritual and symbolism.4 The minaret of the mosque,of which now only the basement and a short stretchof wall remain, is not aligned in the same directionas the mosque itself. It is also not directly connectedwith it and evidently belongs to a different and ear-lier period. Ayverdi and Travlos note this on their plansbut give no explanation. If one checks the orientationof the minaret-cum-wall fragment with a compass itbecomes apparent that it is oriented due east, and wasadded to the southern lateral wall of the church. Whenthe old church was knocked down and replaced bythe mosque properly oriented towards Mecca, for piousreasons the slightly older minaret was respected. Theexcavated foundations of the choir of the old churchin the east and the site of the minaret in the westindicate the size of the church, which must have beena basilica, for it is much too long for a normal cross-in-square plan. As basilicas, with their pronounced east-west axis, are ill-suited for the Muslim prayer, this mustbe the reason why it was knocked down and replacedby a new mosque soon after its conversion.4 1

Now we can be very sure that the Fethiye mosqueof Athens was not the earliest quatrefoil building buta copy of a then fashionable plan, which just a fewyears before had been applied to the very prominentIstanbul Yeni Cami of the Queen Mother Valide Hadice

Turhan Sultan (1666, its foundations were laid in the1590's).42 Apparently the concept of the Fethiyemosque did not come to Athens directly, but indirectlyvia the nearby city of Thebes, where in 1666-67,Egribozlu Ahmed Pasha, brother of the Grand Admi-ral of the Ottoman fleet, Ktse Ali Pasha of Egriboz,had built a monumental mosque using the quatrefoilplan. It disappeared in the nineteenth century, but arelatively detailed picture of it survives on a great iconof its patron saint St. Luke in the cathedral of Thebes.On the icon the saint is depicted full size with a de-tailed panoramic view of "his" city in the background.The icon was painted around 1700; Evliya Qelebi de-scribed this mosque when it was just built.4 3

The connection between the queen mother andThebes is illustrated by the khan she had built in thattown as one of the waqf possessions of her greatmosque in Istanbul. The quatrefoil mosque of Egri-bozlu Ahmed Pasha in Thebes was much bigger thanthe Athens building. It followed a variant of the planused in the mosque of Abdurrahman Pasha in Tosya(1584). In it the two corner domes of the mihrab wallhave been omitted and the half-dome of the mihrabsection shaped more or less to resemble the apse ofthe church. Judging by its picture on the icon in theThebes cathedral, the now vanished mosque of Thebesdominated the town44 (now a dull and featureless place,despite its multi-millenarian history4 5 ), just as themosque of Abdurrahman Pasha dominated Tosya.

Finally, we should say some word about the subse-quent career of the plan. After his epochal ehzademosque, Sinan moved to other plans. He did not comeback to it, aside from its truncated version at Mihri-mah's mosque in Uskfidar, which features three half-domes like the earlier mosque in Oguzk6y and theSiileyman Pasha mosque in the Cairo citadel (earlysixteenth century). Although the plan was used inQankln in 1558 and in Tosya in 1584, we do not knowif it came there via Istanbul, or as a result of the localbuilding tradition. The same plan was also used in theEuropean parts of the empire, although not "through-out the Balkans," as Grabar stated (he could only havehad one example in mind, the misdated Fethiyemosque of Athens, since the prominent example inThebes was unknown when he wrote).

Only one little known example of this mosque typein the Balkans dates from the sixteenth century, andthat is the mosque of Sultan Murad III in the newlyfortified town of Navarino (Pylos) on the southwest-ern peninsula of the Morea. The fortress town of Yeni

�·-�'"·-------�C·---·�·��·----"------··-

118

THE QUATREFOIL PLAN IN OTTOMAN ARCHITECTURE: THE "FETHIYE MOSQUE" OF ATHENS

Anavarin (New Navarino) was built by the Ottomansas an answer to the threat of Western naval attacksafter the disastrous battle of Lepanto in 1571. Thepentagonal citadel, the two heavy coastal batteries, andthe town walls were constructed in 1572-75. In 1576the local Ottoman authorities petitioned the sultanto provide the new town with a masjid in the citadeland a Friday mosque in the walled town. The sultanagreed, and construction of the buildings began soonafter. The correspondence about the building of thecastle and the mosque is contained in the MfihimmeDefters in the Ottoman archives in Istanbul and havebeen partly published by Gfilsfin Tanyeli in theFestschrift for Dogan Kuban. 46 The mosque is currentlybeing restored and investigated by a student fromUtrecht University.

The Navarino mosque is a very provincial versionof the quatrefoil plan, but the relation with the greatIstanbul buildings is evident. The plan found its great-est representatives in Istanbul, where in the post-Sinanperiod it was used for the initial concept of the YeniCami (begun in the 1590's by Valide Safiye Sultan, butcompleted in the mid-seventeenth century by ValideHadice Turhan Sultan) and the famous "Blue mosque"of Ahmed I, begun in 1609. 4 7 That the plan remainedpopular and was regarded' as typically Ottoman canbe seen in the Houmt souk on the Tunisian island ofDjerba. There, in 1640, the mosque of the Strangers(Jamaa Gureba) was enlarged. It was a Hanafi mosquefor the Ottoman garrison; the bulk of the island'spopulation was Ibadite. The older part of this mosquewas built in the local tradition: seven naves run paral-lel to the qibla wall, covered by a flat wooden ceilingwhich is carried by six rows of four thin marble col-umns. The 1640 annex faithfully follows the quatre-foil plan, giving a typically Ottoman stamp to the veryun-Ottoman architectural and human environment ofDjerba.

In 1659-60, the so-called mosque de la P&cheriein the city of Algiers was built by the Janissary corps(ocak) of the town. It is clearly inspired by the quatre-foil plan, but is even more provincial and un-Ottomanin execution than Murad III's mosque in Navarino.4 8

The influence of the Istanbul Yeni Cami, then justcompleted, is evident: This is even more the case inanother North African building from the Ottomanperiod, the mosque of the national saint of Tunisia,Sidi Mahrez, in the north of the old town of Tunis. Itwas conceived in grand style in 1692-97; the Muradiregency of Tunis under Muhammed Bey evidently

wanted to make a statement.4 9 It is the most "Otto-man" building of the entire Maghribi architecturaloutput. The quatrefoil plan was now generally acceptedas being the personification of Ottoman architecturein the way Grabar phrased it.

The Thebes and Athens buildings suggest that thecompletion of the Yeni Cami in Istanbul in 1664 pro-vided the major impetus for its spread. When in 1767an earthquake knocked down the old Fatih MehmedII mosque in Istanbul, which originally had only onehalf-dome over its mihrab, it was almost inevitable thatit would be replaced by a grandiose quatrefoil build-ing.5 0 This in turn inspired Muhammad 'Ali to havehis even bigger mosque on the citadel of Cairo con-structed according to a plan that had become the mostrecognizable symbol of imperial power. For the samereasons it was also used in the mid-nineteenth cen-tury when the main mosque of the great Syrian cityof Homs was reconstructed. The plan had its originin the creative reformulating of a combination of in-fluences, in which local Anatolian ones definitelyplayed a role. Not Athens, but Anatolia was the cradleof one of the most successful and expressive of allbuilding concepts in Islamic architecture.

University of UtrechtUtrecht, The Netherlands

NOTES

Author's note: This article is one of the products of an academicyear at Harvard University under the aegis of the Aga Khan Pro-gram for Islamic Architecture. The research in the Turkish ar-chives and libraries and in the field which forms its base wassponsored by the Netherlands Organization for the Advancementof Scientific Research (Z.W.O./N.W.O) in The Hague, and theDeutsche Forschungsgemeinschaft (DFG) in Bonn.

1. Ulya Vogt-G6knil, Trkische Moscheen (Zurich, 1953), p. 38.2. Behget Unsal, Turkish Islamic Architecture in Seljuk and Otto-

man Times (London, 1959), pp. 24-28.3. For the old Fatih mosque, see E. H. Ayverdi, Osmanh Mima-

risinde Fatih Devri (Istanbul, 1973), pp. 356-406. For theEdirne and Tire buildings, idem, Osmanh Mimarisinde elebive II. Murad Devri (Istanbul, 1972), pp. 37 7 -8 1 and 542-48.The dates for the waqfiyya, but not the completion of theEdirne and Tire buildings, are given in the foundation in-scription. Also see inci Aslanoglu, Tire'de Camiler ve Of Mescit(Ankara, 1978).

4. For the itinerary of the 'Iraqayn campaign, see Ismail HamiDaniamend, Izahh Osmanh Tarihi Kronolojisi, vol. 2 (Istanbul,1971), pp. 167-81; or the beautifully illustrated Beydn-imenfizil-i sefer-i 'Iraqeyn-i Sultan Sleyman Han of MatrakclNasuh, ed. Hfiseyin Yurdaydln (Ankara, 1976), which wascompleted in 1537.

1�I� �I� � _ I�_

119

MACHIEL KIEL

5. Metin S6zen, "Hacl Hamza'daki Turk Eserleri," in AnadoluSanatz Araytzrmalan, vol. 2 (Istanbul, 1970), pp. 113-34. BeforeS6zen, the Oguzk6y mosque was noted by Mahmud Gologluin his little-known work Paphlagonia (Istanbul, 1948).

6. Published as 438 Numaral Mliuhdsebe-i Vildyet-i Anadolu Defteri(937/1530), vol. 2, Bolu, Kastamonu, Kengiri ve Koca-ili Livalan(Ankara, 1994). For the essential data on the life of CobanMustafa Pasha, see Mehmed Sfireyya, Sicill-i 'Osmdni, vol. 4(Istanbul, 1311 H.), p. 37 2. For his magnificent Gebze build-ings, see Godfrey Goodwin, A History of Ottoman Architecture(London, 1971), pp. 189-91. Many more details are includedin Fatih Mfiderrisoglu, "BAni Goban Mustafa Pasa ve birOsmanlh sehri Gebze," in Vakzflar Dergisi 25 (Ankara, 1995),

pp. 67-124.7. Goodwin, History of Ottoman Architecture, p. 175, n. 155; Unsal,

Turkish Islamic Architecture, pp. 24-28, states that the mosquewas given its present form by Alafiddevle. He does not seemto have visited it, since he mistakes the four lateral vaultsfor half-domes because of the thick plaster used in theirrestoration. They were originally eliptic vaults supported onsquinches and the sustaining arches of the central dome.

8. For examples in Central Asia, see Goodwin, Histor, of Otto-man Architecture, n. 156, where he mentions the ninth-cen-tury Degaron mosque in the village of Hazara near Bukhara,published by Yuri Yaralov, "Architectural Monuments ofMiddle Asia of the Eighth to Eleventh Centuries," in FirstInternational Congress of Turkish Art (Ankara, 1959), p. 367,and a few less obvious examples.

9. Goodwin, History of Ottoman Architecture, pp. 177-78.10. Ibid., p. 207. For the Fatih Pasha mosque in Diyarbaklr, see

Metin S6zen, Diyarbakzr'da Tfirk Mimarisi (Istanbul, 1971),pp. 65-69 and 260-63, with older literature in French,German, and Turkish.

11. Heinz-Jfirgen Sauermost and Wolf-Christian von der Mfilbe,Istanbuler Moscheen (Munich: Bruckmann, 1981), pp. 116-17. I was unable to find any further literature on the mosqueof Piri Pasha at Hask6y, and Professor Baha Tanman ofIstanbul University has since told me that he and his col-laborators had made a special study of the area, and thatthe mosque of Piri Pasha in fact no longer exists. It is men-tioned in Hfiseyin Ayvansarayl's Hadikatfi'l-cevami', 2 vols.(Istanbul, 1281 [1864-65]), 1: 308, and by Tahsin Oz, inIstanbul Camileri, 2 vols. (Ankara, 1962-65), 1: 115. Thereis a mosque of Kececi Piri in the Hask6y-Sfitlfice district onthe Golden Horn, but the founder was no pasha and hismosque is built entirely of wood and has no domes whatso-ever. Godfrey Goodwin saw the eighteenth-century lengerhane(factory), a building in poor repair which was indeed usedas a depot. It still exists and has a central dome on four piers,four small domes at the four corners, but instead of realhalf-domes the four remaining sections are covered withcradle-vault-like groin vaults. It is exactly the same buildingtype as the Mevlevi-hane of Kilis and (Gazi) Antep in southernTurkey and the nearby Halep in Syria. The Haskby lengerhanehas, on the basis both of style and of construction techniques,to be dated ca. 1740. It is, according to Tanman, definitelyan eighteenth-century work. In a letter of November 2001,Godfrey Goodwin wrote that he agreed with the identifica-tion of Piri Pasha's mosque proposed by Tanman. When hevisited the building in the 1960's it was very dirty and filled

with tobacco and therefore unsurveyable; and when offeredthe lengerhane option he accepted it wholeheartedly. Themosque of Piri Pasha in Hask6y, Istanbul is thus a simplemisidentification and is definitely not the link betweenDiyarbaklr and the Istanbul ehzade mosque. The lengerhanewas restored a few years ago and is now a part of the RahmiKoc Sanayi Mfizesi. There remains the Fatih Pasha mosquein Diyarbaklr, the direct source of inspiration for the qua-trefoil plan, once the Piri Pasha in Hask6y has been removedas intermediary.

12. Ekrem Hakkl Ayverdi, Osmanh Mi'mdrisinde: Celebi ve I. MuradDevri (Istanbul, 1972), pp. 136-50, with plans, section, theinscriptions, and numerous photographs. It is thirty yearssince Ayverdi wrote about this majestic building, and it isstill in a shameful state of neglect.

13. For the Edirne Eski Cami, also see Ayverdi, Celebi ve II. MuradDevri, pp. 150-60. During the recent restoration (in 2000,still not completed) the lovely and colorful interior of thismosque, with its fine wall paintings in the style of the Otto-man baroque, was totally destroyed and replaced by a dulland lifeless decoration in two colors: coffee with and cof-fee without milk. It is difficult to say which is worse, Greekneglect as in Dimetoka, or Turkish "restoration" as in Edirne.For the tree-ring data, see Peter Kuniholm, "Dendrochro-nologically Dated Ottoman Monuments," in A Historical Ar-chaeology of the Ottoman Empire, Breaking New Ground, ed. UziBaram and Lynda Carrol (New York, 2000), pp. 93-135.

14. Oktay Aslanapa, Osmanli Devri Mimarisi (Istanbul, 1986),Diyarbaklr, Fatih Pasha Camii (of Biylklh Mehmed Pasha,1515-20), pp.152-54; Elbistan (Dulkadirli Ali Bey, 1515-22),pp. 154-55.

15. Ibid., pp. 182-83.16. Aptullah Kuran, Sinan: The Grand Old Master of Ottoman Ar-

chitecture (Istanbul and Washington, D.C., 1987), pp. 67-68.17. Ibid. In his notes he refers to S6zen's important work, Diyar-

bakzr'da Turk Mimarisi, pp. 65-69.18. Ulya Vogt-G6knil, Sinan (Tfibingen and Berlin, 1993), p. 22.19. Ibid., n. 10.20. Dogan Kuban, Sinan's Art and Selimiye (Istanbul, 1997), p.

68.21. M. Tasdemir, "Elbistan," Tiirkiye Diyanet Vakfz Isldm Ansiklo-

pedisi, vol. 11 (Istanbul, 1995), pp. 1-3.22. Refet Ysnanc, Dulkadzr Beyligi (Ankara, 1986), pp. 121-22.

For a list of Alafiddevle's buildings, see ibid., pp. 119-53.The waqfiyya of one of Alafiddevle's mosques in Elbistan ispreserved in the archives of Vakiflar Genel Mfidfrligfi inAnkara and is dated 1501.

23. For a short description and some very good photographsof the Tosya mosque, see Aslanapa, Osmanlz Devri Mimarisi,pp. 300-1. The article "Cankln" in Isldm Ansiklopedisi statesthat the mosque was built in 1557-58; in his History of Otto-man Architecture, Goodwin (p. 177) points to difficulties: "Aproblem is the mosque at Cankin which is said to have beenbegun in 1522 and completed in 1558, a quite inexplicablelength of time. It is a centralized mosque with four semi-domes which are slightly pinched because they are set abovepointed and not semicircular arches. It would look if at firstthey were, or were intended to be, vaults, in which case themosque would be a logical and grander successor to that ofElbistan.... The damage by earthquake was such that the

_C� ��sl� I_

120

THE QUATREFOIL PLAN IN OTTOMAN ARCHITECTURE: THE "FETHIYE MOSQUE" OF ATHENS

repairs in 1302/1884 were partly a rebuilding...." Goodwintook the two different dates from a Turkish work, cited asYalqn, Cankzrn, p. 10. The date of the major repair is inanother inscription on the mosque. Whether it is 1522 or1558 (which fits better) is irrelevant for our theme. Afterthe earthquake the building was not changed fundamen-tally but remained a quatrefoil plan.

24. Ginkut Akin, "Dogu ve Gfineydogu Anadolu'daki TarihselEv Tiplerinde Anlam," Ph.D. diss., Istanbul Technical Uni-versity, 1985. Also in his Asya Merkezi mekdn Gelenigi (Ankara,1990), pp. 194-98 (with other examples outside Turkey).Nezih Bagelen, Bir Dogu Anadolu Kiyinn Kiiltirel GemisiUzerinde Arastzrma: Onar Dede mezarlzhg ve ad2 bilinmeyen birTiirk kolonizat6rii seyh Hasan Onar (Istanbul: Arkeoloji ve SanatYaymlari, 1983); and long discussions with him in Istanbul.

25. For examples in northern Pakistan, see Peter A. Andrewsand Karl Jettmar, Sazin: A Fortified Village in Indus-Kuhistan,Antiquities of Northern Pakistan (Mainz: Ph. v. Zabern Ver-lag, 2000), p. 152, with many detailed drawings and plans.I would like to thank Dr. Andrews of Cologne University forsharing with me his vast and thus far unpublished knowl-edge of the spread of the Kirlangic dome.

26. Hakim Karpuz, Tirk islam Mesken Mimarisinde Erzurum EvIeri(Ankara, 1984).

27. A. Xyngopoulos, "Ta Vyzantina kai Tourkika Mnimeia tonAthinon," in G. Sotiriou, A.K. Orlandos, A. Xyngopoulos,Heureterion ton mesaionikon mnimeion tis Ellados, vol. 2 (Athens,1929), pp. 116-20; and vol. 3 (1933), p. 2 2 9 . Also see Orlan-dos, "Ergasia anastiloseos Vyzantinon mnimeion," in ArcheionVyzantinon Mnimeion tis Ellados, vol. 3 (1937), pp. 203-6.

28. Semavi Eyice, "Yunanistan'da Tfrk Mimari Eserleri," in Tiirki-yat Mecmuasz 11 (1954), pp. 157-64.

29. loannis Travlos, Poleodomiki exelixis ton Athinon, 2nd ed.(Athens, [1960] 1993), pp. 181-82: "The Fethiye (the Con-queror) . . . is as it appears, one of the first mosques builtby the Turks immediately after they had settled in the town."In the same work (p. 181) he gives a plan and photographof the mosque, accompanied by the caption, "Second halfof the 15th century." In his monumental work from 1971,Pictorial Dictionary of Ancient Athens (London: Thames andHudson, 1971), Travlos published an excellent plan of thearchitectural environment of the building, accompanied witha detailed air photograph. Ayverdi (Osmanli Mimarisinde FatihDevnri, n. 23), writing much later, made no use of either ofthese excellent presentations.

30. Ayverdi, Osmanlz Mimarisinde Fatih Devri, 3: 49-55.31. Aslanapa. Osmanlz Devri Mimarisi, p. 103.32. Nusret Cam, Yunanistan 'daki Tuirk Eserleri, Tfirk Diyanet Vakfi

(Ankara, 2000), pp. 15-19, Athens, Fethiye Camii from 1459,but with a question mark from the author to convey hisdoubts.

33. S. Eyice, article "Atina," in Isldm Ansiklopedisi, vol. 4 (Istanbul,1991), p. 78.

34. The 1506 register is not preserved as a whole. The bulk isformed by T.D. 35 in the Basbakanlik Osmanli Arivi (hence-forth BBOA) in Istanbul. A sizeable fragment was sent inthe 1930's as waste paper to Sofia, where it was not-as in-tended-transformed into pulp but catalogued and preservedin the National Library and recently published by EvangeliaBalta in Archeion Evoikon-Meleton (29.1990/91, pp. 55-

185). Another fragment was catalogued by Ali Emiri in theearly 20th century as Ali Emiri Tasnifi, Bayezid-i Sni 04.The contents of the 1521 register was reproduced in T.D.367 from 1528-30. The 1540 register is T.D. 431, waqfs onp. 1056. The 1570 register is preserved in Ankara and, a moreaccessible copy, T.D. 484 in Istanbul, BBOA. The texts ofthe three tahrirs, showing the number of households ofAthens in facsimile, have been published by M. Kiel, "Cen-tral Greece in the Sfileymanic Age: Preliminary Notes onPopulation Growth, Economic Expansion and Its Influenceon the Spread of Greek Christian Culture," in Soliman leMagnifique et son temps, ed. Gilles Veinstein (Paris: Ecole duLouvre, 1992). pp. 399-424.

35. As late as 1953, Demetrios Sicilianos, in his history of Ath-ens, translated by Robert Liddell as Old and New Athens (Lon-don, 1960), p. 236, still called the building "this fine Byzantinechurch, once the cathedral of Athens under the title PanayiaSotira, is . . . in the corn market of old Athens. After thefall of Athens in 1460 this church was turned into a mosquein honour of Mehmet the Conqueror." In the Greek histo-riography the connection between the Fethiye Mosque andthe church of the Panaya tou Staropazarou was first pointedout by Dimitrios Kambouroglou in his seminal works Istoriaton Athinaion, Tourkokratia (Athens, 1889), and Mnimeia tisIstorias ton Athinaion epi Tourkohratias, 3 vols. (Athens, 1889-96). Dealing with the church/mosque in vol. 2 (pp. 177 and179) of his Mnimeia, Kambouroglou suggests (but does notwrite) that the church was converted into a mosque shortlyafter the conquest of the city in 1456. It was the Orthodoxcathedral of the city because after 1204 the Frankish rulersused the former cathedral, the Parthenon temple, as a Ro-man Catholic church. Kambouroglou's source was the de-tailed travel account of Spon and Wheler from 1675-76 (seeinfra, n. 38). One has the impression that the confiscationof the church and subsequent rebuilding as a mosque in1669-70 was locally entirely forgotten.

36. Evliya Celebi, Seyahatname, vol. 8 (Istanbul, 1928), p. 450.I also checked Evliya's autograph copy, Bagdat K6skfi 303,in the Topkapi Palace Library, but there are no importantdifferences.

37. Paris, Bibliotheque Nationale, Supplement turc 1027, fols.123b-124a. I am currently working on an edition of this raretext.

38. J. Spon and G. Wheler, Voyage d'Italie, de Dalmatie, de Grace,et du Levant (Lyons, 1678); separately published, with otherdetails as George Wheler, A Journey into Greece in Company ofDr. Spon of Lyon (London, 1682). Especially the first versionwas a bestseller in its time and was translated into Dutch,German, Swedish, and other languages. The quotation isfrom p. 391.

39. Ankara, Vakiflar Genel Mfdfirlfiig Arsivi, Haremeyn I, Defter734, slra 141, fols. 137b-139b. I am currently preparing thisdocument-interesting for its information on Islamic lifein Athens-for publication.

40. Most important is that the first light of the day, symboliz-ing the rising of Christ from the dead, falls on the altar wherehis mystery is celebrated in the mass. For details, see LeonidOuspenskij, "Symbolik des Kirchengebdudes," in Symbolik derReligionen, ed. F. Herrmann, vol. 10 (Stuttgart 1962), pp. 56-62.

121

·--- "C· �iiSC-n �II-·-

MACHIEL KIEL

41. A different opinion on the origin of the Athens Fethiyemosque is expressed by Dimitris N. Karidis in his disserta-tion, "Poleodomika ton Athinon tis Tourkokratias," AthensUniversity, 1981, p. 275; he states that there is no evidenceof a Greek Orthodox cathedral in the Athens Corn Market;Spon and Wheler had merely confused it with another churchnearby. The church ruins underneath the Fethiye mosqueare from a much older church, which was ruined centuriesbefore. The site, at the crossing of two of the most impor-tant streets in Ottoman Athens was chosen for the deliber-ate urbanistic purpose of having a dominating structure atthis important locus. As these streets no longer exist (theyhave been dug away by archaeologists), Karidis gives detailedreconstruction plans for the original site. He explains thenon-aligned minaret as a survival of an older, perhaps six-teenth-century mosque, and regards the story of the dervishtraveler from 1790 as pious folklore. It could also be arguedthat the story heard by Spon and Wheler echoes the localmemory that a church had long ago existed on the site wherethe mosque was about to be built. This theory has the his-toriographic tradition of more than a century against it, buthas the advantage of doing away with the painful event forthe local Christians of losing their cathedral church. Thatthis event left no trace in the memory of the Athenians andwas "discovered" by Kambouroglou as late as the end of thenineteenth century, supports Karidis's theory. The theorydoes not affect the conclusion that the mosque dates fromthe 1660's.

42. For this grand building, see Sauermost and von der Mfilbe,IstanbulerMoscheen, pp. 200-5; "istanbul Yeni Cami ve HfinkarKasri," Vakzflar Genel Midiirligi2 Yayznlart (Istanbul, [ca.1985]), a publication of the waqfiyya of this great founda-tion with a vast number of illustrations. Also see DoganKuban, "Yeni Cami Kfilliyesi," in Istanbul Ansiklopedisi, vol.7 (Istanbul, 1994), pp. 464-67, and Lucienne Thys-$enocak,"The Yeni Valide Complex at Emin6nfi," Muqarnas 15 (1998):58-70.

43. For the icon in Thebes and the great mosque of Ahmed Pa-sha, see M. Kiel, "The Icon and the Mosque, Orthodox Chris-tian Icon Paintings as a Method to Retrieve the Lost OttomanArchitectural Heritage of the Southern Balkans," in Aptullah

Kuran iin Yazzlar, ed. Cigdem Kafescioglu and LucienneThys-Senocak (Istanbul, 1999), pp. 223-32; published in En-glish as Essays in Honour of Aptullah Kuran (Istanbul, 2000),pp. 230-45.

44. Ibid., p. 245.'45. It was in Thebes that Europe got its name when, according

to ancient Greek mythology, a Phoenician princess from Tyrenamed Europa was abducted to Greece by Zeus, resultingin a Phoenician intervention under her brother Kadmos.Most philologists regard the name Kadmeia, of the acropo-lis hill of Thebes, as being of Semitic origin. The strong-hold was built by Kadmos, after he became king of Thebes.The name Kadmos is taken as proof of a Semitic Phoenicianinvasion of Greece in prehistoric times.

46. Gfilsim Tanyeli, "Bir osmanli Kale-Kentinin Yaplm: AnavarinOrnegi," in Prof Dogan Kuban Armaganz, ed. Z. Ahunbay, D.Mazlum, K. Eyfipgiller (Istanbul, 1996), pp. 85-93.

47. Goodwin, History of Ottoman Architecture, pp. 342-49. Sauer-most and von der Mfilbe, Istanbuler Moscheen, pp. 206-14. Amass of details on the construction of this mosque can befound in Zeynep Naylr, Birinci Ahmet Klliyesi (Istanbul, 1975);and Howard Crane, ed. and trans., Risale-i Mi'mariyye: AnEarly-Seventeenth-Century Ottoman Treatise on Architecture(Leiden: E.J. Brill, 1987).

48. For photographs and description of this building, see GeorgeMarcais, L 'Architecture musulmane d'Occident, Tunisie, Algirie,Maroc, Espagne et Sicile (Paris, 1954), pp. 430, 433-34. Moredetailed is R. Dokali, Les mosqudes de la periode turque Alger(Algiers, 1974).

49. Slimane Zbiss, "Des aspects tunisiens de l'art turc," in Com-munications of the First International Congress of Turkish Art (An-kara, 1961), pp. 377-81, and photographs on pl. CCXCIV;idem, "Les monuments religieux dans l'architecture turc deTunisie," in Atti del Secondo Congresso internazionale di arte turca(Naples, 1965), pp. 395-96. See especially Ahmed Saadaoui,"Une architecture turque en Tunisie: La mosquee de Mu-hammad Bey (1675-1696)," IBLA: Revue des Belles Lettres Arabes185 (2000-1): 3-31.

50. Sauermost and von der Mfilbe, IstanbulerMoscheen, pp. 214-24; Goodwin, History of Ottoman Architecture, pp. 394-95.

Il--r�-------rr�r�·lIOLIIICLI·I� I pi -

122