Bell's discussion of Near Eastern domes relies on some of her earlier observations of such features, the product of her investigations of ecclesiastical architecture in Anatolia, much of which appears in the 1909 work The Thousand and One Churches.116 She brings in the evidence from the western part of the Near East, mainly the coastlands of Anatolia, where architects began to consider more seriously solutions for setting a spherical roof elegantly over a square plan. They accomplished this by introducing spherical pendentives, these basically consisting of curved triangular pieces of brick or stone masonry that rise up from the corners of a square substructure, transforming its angles into a circle upon which the dome can be set. Bell conjectures that the earliest examples of domes with pendentives date back even before the Age of Constantine (i.e., earlier than the third century CE), but that they became prolific in Western Anatolia and Syria in the Byzantine period of the sixth century and found their truest and most magnificent expression in the form of the massive dome of Justinian's Santa Sophia in Constantinople.117 Bell also refers to an often cited dome at Jerash (in present-day Jordan) as a very early domed space with pendentives from Syria, although she hesitates to identify it with the pre-Christian Era; rather, she believes that it should be contemporary with Justinian's examples from Constantinople.118
Fig. 3.18 Bell’s photograph of the remains of a groin vault, located in the northwestern corner of Room 141 in the East Annex. The corbelled springer in the corner is still preserved, although the rest of the groin has fallen in.
Fig. 3.19 Bell’s photograph of the groin vault in the north-east angle of Corridor 28, with its plaster still intact. These types of intersecting vaults are rare at Ukhaidir, only appearing in eight examples in the palace and adjoining Eastern Annex.
In contrast to domes set on pendentives, Bell describes what she refers to as a more primitive form of dome construction, this prevalent among the inland Anatolian structures she had investigated, and dating to the fourth and fifth centuries.119 Quite simply, the rectangular base of the space below was transformed into an octagon by placing horizontal brackets (‘corbel stones’) across the corners, and from this setting it was possible to construct a circular dome.120 This type of dome construction remained quite prevalent in inner Anatolia and elsewhere in the inland parts of the Near East for several centuries.121
In addition to her examination of the Western development of the dome, Bell endeavoured to track the history of the dome in the East. In Persia, the Sasanians had arrived at their own solution for constructing a dome over a rectangular base. This involved using squinches, these being arched niches set over the corners of a room, serving to transform the angles into curves and thus allowing for the placement of the dome above.122 Squinches enabled Sasanian builders to throw their domes over a span of up to 16 metres, a technique that can be observed at the site of Firuzabad, the earliest of the Sasanian palaces, and at Sarvistan, which was in Bell's time widely regarded as a good example of a Sasanian structure from the fifth century CE.123 Bell also conjectured that several of the spaces within the ‘small’ and ‘large’ palaces at Qasr-i-Shirin also presumed to be of Sasanian date (the Palace of Khosrow and the Chehar Qapu), although no longer intact, had been domed with the use of squinches and had spanned up to 16 metres.124
Fig. 3.20 Bell’s photograph of the south-western corner of the fluted dome in Room 4. The dome may have been originally furnished with a circular aperture at its apex. One can see the somewhat primitive way in which the dome has been set on brackets in the corners of the square room rather than with squinches or pendentives, which would have transformed the angles of the room into the curves of the dome more elegantly.
When Ukhaidir's evidence for domes was considered in light of this amassed data, Bell noted that it could not be compared easily to the domes of the palaces to the east. While it is certainly true that Ukhaidir's architects were familiar with squinches, they never used them to construct true overhead domes. Rather, squinches were only employed to resolve angles between barrel vaults or corners.125 One can observe such use, for example, in a corner of the second-storey gallery (no. 134) overlooking Court A at the north end of the castle (Fig. 3.21).126 There are also examples of squinches in the southern arcade of Ukhaidir's mosque (no. 11), where they occur in the corners of semi-domes set between transverse ribs. One particularly well-preserved squinch in the mosque arcade is decorated with a fluted stucco pattern, flanked by shallow pointed niches (calottes), while above are concentrically recessed rosettes and crenellated brick patterns.127 So there was certainly knowledge of the use of the squinch, whose idea the Ukhaidir architects may have borrowed from the architectural traditions of the East, but they do not seem to have been confident enough in their own skills to use the squinch to build true domes – certainly none with the wide spans of the monumental Eastern palaces.128
Of the domes at Ukhaidir, none exhibits the technique emanating from the West, described earlier, in which spherical pendentives rise from the corners of the angled spaces, allowing for the placement of a circular dome above. The Ukhaidir fluted dome depended instead on horizontal slabs placed obliquely across the corners of the square rooms, the other technique that we know emanated from the western part of the Near East but which is seen as a more primitive type of dome placement and figures prolifically in the earlier fourth- and fifth-century churches of inner Anatolia.129 This feature might tempt one to postulate an early date for the domes of Ukhaidir as well, and yet geographical distance is conceivably the explanation: perhaps Western techniques had not yet reached the architects of Iraq.130 In the end, the presence of the domes and squinches did not help Bell to fix a more precise date for the construction of Ukhaidir, but in the process of discussing these features, their points of origin and their specific manifestations within the palace, Bell drew attention to the complex, multi-directional nature of the inspiration and influences working upon its architecture. The recognition of Ukhaidir's multifarious character is quite significant; scholars both contemporary with and later than Bell frequently emphasize this diversity in the art and architecture of the Early Islamic period.131
Masonry Tubes
One other item relating to Bell's observations of the vaulting at Ukhaidir concerns somewhat unusual but practical features that occur several times, and which Bell describes as masonry tubes. They are hollow, vaulted galleries that run between adjacent chambers with parallel barrel vaults of the same height, as well as between vaults and straight walls.132 The openings of these tubes are located high up in the spandrels between vaulted spaces, and their ends can often be seen in the façades of open courts, such as those of the iwan groups within the baits on the main floor of the palace (Fig. 3.22), and also in the façades of the iwan groups facing the open courtyard on the third storey of Ukhaidir's gatehouse.133 The function of these hollow tubes would have been to relieve the enormous weight of the stone masonry of the vaults, although Bell suggests that they might also have served to keep the rooms cool by providing a belt of unheated air along the vaults.134 For parallels to these tubes, Bell mentions their early presence at the Parthian site of Hatra, where they appear in some of the tombs.135 On the other hand, Bell sighted masonry tubes at the thirteenth-century Khan Khernina above Tekrit and noted that this architectural tradition must have carried on for quite some time among Islamic builders.136 She also suggests that these openings formed an essential part of the Islamic façade, transforming, for example, into windows and niches on either sides of arches in the façades of the mosque of Ibn Tulun in Cairo and the mosque of Abu Dulaf at Samarra.137 In all, the presence of masonry tubes connected Ukhaidir to earlier supposed Sasanian antecedents, but to Bell it also underlined their continuing ubiquity in the early Islamic architecture of Mesopotamia, a period of which she was slowly taking greater notice as she considered Ukhaidir's architectural features as a whole and noted the great number of parallels they had to the structural attributes of that time.
Fig. 3.21 Squinch in on
e corner of Gallery 134 on the second storey of the palace’s northern gateway block.
Ukhaidir's Mosque
The north-western corner of the Ukhaidir palace – with its unique arrangement of open court and arcades decorated with stucco ornament, and doors framed by elaborate cusped patterns – marked this section as special and distinctive. Bell's diary and letters of 1909, when she first visited Ukhaidir, provide no conjecture about this area, but already in her first scholarly publication on Ukhaidir, which appeared in early 1910, she raises the possibility that this might have been a mosque.138 Significantly, the German scholar Ernst Herzfeld was quite struck by Bell's suggestion, reasserting this identity in his own article later in 1910, a piece on the date of the palace of Mshatta in the western Syrian desert. A similarly placed chamber in the Mshatta palace had led Herzfeld to conjecture the presence of a mosque there as well, thus suggesting an Islamic date for both desert palaces.139 The only difference was that a niche in the south wall of the Mshatta chamber – ostensibly a mihrab – further strengthened its identity as a mosque, whereas this detail had not been reported at Ukhaidir.140 In Amurath to Amurath, Bell returned to the issue of the identity of the north-western sector of Ukhaidir, offering a mosque as a possible function and citing Herzfeld's own findings, as well as his agreement with this idea.141 But it seems that she was still undecided on the matter at this point, for alongside her suggestion of an early Islamic date for the time of Ukhaidir's construction she also offered an earlier Sasanian date.142
Not until Bell's final report on Ukhaidir, which appeared in 1914, did she put in print her absolute certainty about the matter. The reason: in the spring of 1910, Bell asked a French architect by the name of Henri Viollet, who was about to embark on a trip to Mesopotamia, to visit Ukhaidir and, while there, clear away the rubble from the centre of the south wall of the cloistered hall to see whether he could find a mihrab niche.143 From one of Bell's letters to her mother, dated 5 January 1911, we know that it was while dining with Viollet and his wife in Paris upon the Frenchman's return from Mesopotamia that he told her news of his investigation at Ukhaidir. He had indeed found a concave mihrab in the place where she had told him to look (Fig. 3.23)! Bell's suggestion that this was a mosque and that Ukhaidir was Islamic in date was now firmly secured.144 Interestingly, this was not the first time Bell had made a positive identification of a mosque. In the course of Ramsay and Bell's investigations of an early antique church at Binbirkilise in Anatolia in 1907, Bell recognized that a stepped platform in the church – previously regarded as a pulpit – was actually an Islamic minbar, and she then cleared away the wall beside it to reveal the mihrab. These features made it evident that the church had been converted into a mosque at some point. We are told that both Ramsay and the local Muslim workmen were deeply impressed and delighted by this discovery.145
The Proposed Date of Ukhaidir and Its Builder
While several scholars postulated dates for Ukhaidir after its details had been published, Bell was among the first to submit the most plausible date, in light of her own investigations and careful research. In his later report, Creswell details the various deliberations over Ukhaidir's dating.146 We need not repeat these data here, save to remark that many scholars, including L. Massignon and M. Dieulafoy, continued even into the 1920s to argue strenuously for a Sasanian sixth-century or early seventh-century date. Bell, meanwhile, had already insisted in 1914 that it must be Islamic and thus be dated after the Hijra (622 CE), given that one sector of the palace functioned as a mosque. Moreover, the finding of a concave mihrab suggested a date after 709 CE, when the first mihrab of this type appeared in the mosque at the prophet's home of Medina.147
Fig. 3.22 South side of central Court B in one of the baits along the eastern side of the palace, showing masonry tubes beside the central arched doorway into an iwan room (no. 48). Many of these masonry tubes likely would have been covered over with plaster.
It remained for Bell to decide precisely when in the Islamic period Ukhaidir's construction and occupation should be placed. Her copying and translation of an Arabic graffito, which was discovered in the passageway between Rooms 44 and 45 in the palace, offered few clues.148 Assisted by two German Arabists, Bernhard Moritz and Enno Littmann, she determined the graffito to have been written around 1369–78 CE. The inscription reports the use of the well at Ukhaidir but does not give any indication as to the date of the structure's creation or its original owner.149
Ultimately, Bell appears to have stood by a date early in the Abbasid period, around the middle of the eighth century CE. Her reason for not placing the complex any later concerned the nature of the arches. While most were slightly pointed, a few examples were round, hearkening back to an earlier Sasanian tradition. This differed from the arches at Samarra, confidently dated by Herzfeld and others to the Abbasid period of al-Mansur in the late eighth century, which are all of the pointed variety. Ukhaidir's architects, therefore, had not adopted the widespread tradition of pointed arches, and thus had to have been designing them at a slightly earlier date.150
As for the identity of the individual who built Ukhaidir, Bell looked to the writings of the historian Yaqut, who mentions that a certain Isa ibn Ali ibn Abdullah, the great-uncle of the Abbasid caliph al-Mansur, had been responsible for demolishing a building in the desert called Qasr al-Muqatil and then subsequently rebuilding it.151 She proposed that the newly rebuilt Qasr al-Muqatil was her Ukhaidir, and that its construction would have occurred around the middle of the eighth century (c.750 CE).152
Creswell had a different idea about the castle's builder and about the date. He noted, like Bell, that some of the architectural features, such as the concave mihrab, place its construction after 709 CE. He agreed with Bell that the pointed arch was not fully established at Ukhaidir, whereas it was at Samarra by around 849 CE. Thus, Ukhaidir must be earlier than this date. Creswell remarks that historically, the Abbasid caliphs did not lead the semi-nomadic lives of their Umayyad predecessors, who built residences for themselves on the Syrian side of the desert. On the contrary, they were town-dwellers and resided, at least after 764 CE, at Baghdad. Nevertheless, it was conceivable to consider other individuals from this general time period, and in this light Creswell proposes that Isa ibn Musa (ibn Ali ibn Abd Allah ibn Abbas), the nephew of as-Saffah and al-Mansur, was the builder of Ukhaidir.153 Once the heir apparent to al-Mansur, Isa received a generous financial compensation for renouncing his claim to the throne and retired from public life around 775 CE. It is said that Isa retired to his estates, where he lived in complete isolation, ‘only riding to Kufa once a week, during two months of the year to attend Friday prayers’.154 As Creswell argues, Ukhaidir would fit Isa perfectly, as the palace could only have been built by a man with such wealth, and Isa is the only Abbasid prince known to have lived in isolation.155 Furthermore, Ukhaidir's distance from Kufa is only around 80 km, which could have been covered in two stages, especially if one considers the use of the site of Khan ‘Atshan as a resting post, given its location almost exactly halfway between Ukhaidir and Kufa.156
The precise identity of Ukhaidir's builder has not yet been confirmed, although many scholars continue to follow Creswell's highly plausible scheme and date.157 Significant, however, is a different reconstruction, presented by Barbara Finster and Jürgen Schmidt. In the 1970s they undertook survey work in the desert east of Kerbela and probed in particular the ruins of Tulul al-Ukhaidir, a site only about 2.5 km to the north of Ukhaidir.158 Following the earlier work of Werner Caskel,159 their postulated reconstruction identifies Tulul al-Ukhaidir as Qasr Bani Muqatil, which was first built in the middle of the sixth century CE.160 In the later Abbasid period, around 762 CE, Isa ibn Ali, the uncle of as-Saffah, pulled down the castle and rebuilt the new Qasr Muqatil at Ukhaidir.161 One can see by this reconstruction that it nicely concurs with Bell's own attribution of Ukhaidir to Isa ibn Ali and her proposal that it was the rebuilt Qasr Muqatil, even if she knew nothing about the remains of the ea
rlier Tulul al-Ukhaidir nearby. As remarked above, Ukhaidir's precise date and identity still cannot be established, but it is worth seeing that Bell's guesswork on these important issues falls closely in line with that of more recent scholars.
Fig. 3.23 Bell’s photograph of the southern side of the mosque, the vaulting almost entirely fallen except for the south-western and south-eastern corners. In the centre of the south wall further down, peeking just above the pile of debris, is the semi-dome of the mosque’s mihrab.
Fig. 3.24 The eastern façade of Ukhaidir’s outer enclosure wall, with its blind arches set between semi-round and round towers. Bell’s shadow figures prominently in the foreground centre of the photo.
Assessment of Bell's Architectural Study of Ukhaidir, and Concluding Remarks
More will be said about the overall reception of Bell's work on the palace and mosque of Ukhaidir once her discussion of this complex as well as her consideration of the origins and evolution of the early Islamic palace in their entirety is discussed in Chapter 5. For now, suffice to say that her analysis of Ukhaidir's features such as vaults and domes, and her positive identification of the complex's mosque, was generally well received by many of her scholarly peers. Reuther, although producing his own work on Ukhaidir, acknowledged Bell's research of the site, including her correct identification of a mosque,162 as did Herzfeld, who included her mosque proposal in his own brilliant article on the development of early Islamic art and architecture and the dating of Mshatta.163 Slightly later, Creswell would himself visit the desert castle four times between 1930 and 1936 and produce his own detailed measurements, plan and photographs, all of which appeared as a full description and analysis in the second volume of his exhaustive Early Muslim Architecture.164 Creswell's description would also appear, in an abridged format, in his A Short Account of Early Islamic Architecture.165 Both would essentially become the standard references on Ukhaidir and are to this day the sources most frequently cited by scholars and students who have an interest in the desert castle and its place within the development of the architecture of the early Islamic period. Yet despite the publicity afforded to Creswell, the details he offers about Ukhaidir are essentially a synthetic treatment of the work presented earlier by Reuther and Bell. Creswell makes liberal use of Reuther's drawings and reconstructions and recounts the latter's fine points about various architectural forms and their methods of construction.166 From Bell's work Creswell copies her discussion of the domed spaces and groined vaults within the palace,167 her identification and suggested function of the masonry tubes,168 her identification of the palace's mosque,169 and her comparison of the layout and architectural features of Ukhaidir to those from Mar Tahmazgerd in Kirkuk, Firuzabad, Ctesiphon, Qasr-i-Shirin and Sarvistan.170 Bell's erudite observations about pitched brick vaults, domes and pendentives are repeated in other sections of Creswell's work.171 In sum, in its substance and organization, Creswell's treatment of Ukhaidir and his work on early Islamic architecture in general owes a tremendous debt to his predecessors, especially Bell.
Britain and the Arab Middle East Page 14