THE SASANIAN PALACES AND THEIR INFLUENCE IN EARLY ISLAM BY LIONEL BIER

THE RUINS OF PALATIAL BUILDINGS IN AND impression that we have before us a complete have been linked with the Sasanian dynasty since unit. The graphics have prepared the ground for the nineteenth century, but the concept of a statements about the building's symmetrical plan Sasanian palace architecture goes back only six and theories about its function. decades to Oscar Reuther's study in the Survey of This transmogrification of an original survey is PersianArt.' Despite excavations and surveys un- particularly striking at where Watelin uncov- dertaken since then, Reuther's work remains ered in a relatively small area what he described extraordinarily influential. Indeed, most of our as eight "Sasanian palaces." 5 Palaces I and II are impressions about the "Sasanian palace" still de- well known for their rich stucco decoration and rive from this study and particularly from the their elaborate ground plans which suggest a attractive drawings with which he illustrated it. ceremonial function. Even when the plans are Reuther's seminal work has many shortcom- hatched rather than blackened, we have become ings, which were due for the most part to the accustomed to seeing in each a more or less nature of the materials available to him. His complete building. Moorey, who has recently firsthand experience of the monuments he pre- made a fresh study of the Kish excavations, sug- sented was limited to where he exca- gested that Palaces I and II may actually have vated in the late 1920s. For everything else he had been part of a single complex if not a single to defer to the accounts of others; Flandin and building,6 and the published plans are here ar- Coste, for example, and the Dieulafoys, de Mor- ranged in a pastiche as if they were (fig. 1). A gan, and Gertrude Bell. These, in turn, had based variance of some ten degrees indicated by their their Sasanian attributions on traditions embod- north arrows does notpose a significant problem; ied in the works of Arab and Persian authors planswhich showancient buildings oriented dead writing centuries after the fall of the empire. Few north are always suspect, especially in roughshod buildings at that time had been adequately re- surveys, which this one seems to have been. The corded and even fewer excavated. Add to this the published site plan, which is apparently defini- fact that these monuments have yielded virtually tive-although it has no scale-seems to indicate no epigraphic material and it is easy to under- a uniform orientation but in a different direc- stand the problems which Reuther faced in com- tion. They fit well, in any case, in their general piling his study. scale and in the thickness of their outer walls It seems to me that now, some sixty years later, which vary from room to room. What lay in a realistic conception of Sasanian palace archi- between and to the north may have fallen victim tecture continues to elude us, and that this is due to the plow, a common fate for mud-brick build- largely to an oddly uncritical acceptance of the ings, but we are not given the topographic infor- published drawings which in the end are our mation to judge. most important source of information. The older The case of is especially interesting in plans, for example, are so familiar through fre- this respect. The original publication, which re- quent reproduction on an ever smaller scale that mains the basic work, contains awell-known plan they have become almost iconic. They most often (fig. 2) showing the great cruciform hall flanked begin, as in the case of Damghan,2 as line draw- by a rectangular court in the south and a group of ings which clearly indicate the limits of excava- three rooms in the north. 7 The plan first ap- tion and preservation, but in time they are re- peared already blackened and, like all drawings duced to their essentials. In the Survey the walls of in this style, has tended to divert attention from the palace are partially blackened for clarity.3 archaeological problems like the separation of Further along in the rescension the broken edges building phases. There is no indication, first of become less distinct. F. Kimball's reconstruc- all, that the partly sunken structure, which was tion,4 which appears in the following chapter, made of dressed stone blocks rather than the shows a clean edge at the left, adding to the usual mortared rubble, and which is actually 58 LIONEL BIER

oriented differently from the rest of the build- means the parts exposed by Ghirshman-was not ing,8 almost certainly existed before the palace a palace at all but a temple for the worship of was built. Nor is there any indication that the Anahita.' 2 My arguments with him stem from the massive walls defining what Ghirshman called the architectural analogies he made with his frag- "triple " were, as Keall recently pointed out, 9 mentary building at Hajjiabad to the south, which later additions, even though they partially cov- I do not find convincing. But his conclusion is ered the famous floor mosaics. entirely reasonable, especially since the cruci- Ghirshman also published an aerial photo- form hall at Bishapur with its associated rooms graph of the city (fig. 3) showing its grid plan, the and courts lies immediately adjacent to the sunk- river, and the citadel at the mouth of the gorge.' 0 en building at the edge of the great complex, One can see that the entire northeast corner of which was, as A. A. Safaraz proposed, most likely the city was occupied by an enormous enclosure an Anahita temple.' s of some 27,000 square meters, whose southern H. von Gall's theory that the Bishapur mosaics limit and southwest corner are plainly visible. To with their strong Dionysiac flavor alluded to the the east is a depression which represents a great Bacchic pomp borrowed by Shapur from western rectangular court measuring approximately 30 x rulers to celebrate his own military victories over 50 meters. In the centers of three sides are the the Romans' 4 would not contradict a cultic inter- remains of structures that were probably . pretation because Sasanian state religion had From the fourth side a broad corridor (which has from the very beginning a strongly militaristic since been cleared by Ali Akbar Sarfaraz) led to character. If the excavated portion of the build- the excavated western portion of the palace which ing was indeed of a sacred nature, the secular seems to have comprised less than seven percent activities and specifically the audience could have of the whole. been located elsewhere in this vast complex, most Such scrutiny of a well-known photograph puts likelyin one of the iwans that opened on the great the palace of Shapur into a somewhat clearer court. In the same vein, it seems entirely possible perspective and has interesting implications for that if Sasanian Palaces I and II at Kish did the thorny problem of functional interpretation originally belong to the same building, one locale not only of Bishapur but of the Sasanian palaces could have served as an audience hall, the other in general. as a chapel. This interpretation has tended to follow two Perhaps the best example of how architectural often interconnected lines. The first has been to drawings can cloud rather than clarify almost any take the sum total of all the nefarious activities issue is the so-called Imaret-i Khusraw, the palace that would have taken place in such palaces and of Khusraw II at Qasr-i Shirin. That this building make them fit the fragmentary remains. Here we can have played such an important role in the are like the three blind men who describe the architectural history of the region is astonishing elephant variously as a snake, a tree, or a whale, because Reuther's wonderful drawing (fig. 4)'5 depending on which part of the beast we happen on which virtually all discussion has been based is to touch. The second is to see these buildings not a total fabrication. The building, which rose from as palaces at all but as fire temples. Almost all of a great platform, was in a ruined state long before the monuments now thought to have been Sasa- de Morgan came through on his mission scientifique nian palaces have at one time or another been in the 1890s. But he managed to extract a plan seen as temples, and some still are." which showed basically a series of bayts around an Once it is recognized thatwhatwe have come to open court, and an elaborate gate complex preced- think of as a more or less complete building is but ed by colonnades that were doubled at the front.'6 a small portion of one, some difficulties disap- A few years later Gertrude Bell visited the site pear. We know from the Pahlavi inscription on and produced another plan which looked vague- the Kaba Zardasht at Naqsh-i Rustam, for exam- ly like that of her predecessors except that, in- ple, that the king and queen and members of the stead of rows of paired columns, she has a simple royal court made religious sacrifices on a daily iwan hall of narrow proportions.17 Now Reuther, basis, so we can assume that the palaces and who gives no indication of having seen the place, perhaps smaller princely residences like those recognized the inconsistencies of the two surveys uncovered at Ctesiphon contained chapels of and tried his hand, explaining that he has taken some sort. Most recently M. Azarnoush has ar- the liberty to make his own variation on a theme gued that the palace of Shapur-by which he based on the columned buildings at Damghan THE SASANIAN PALACES AND THEIR INFLUENCE IN EARLY ISLAM 59 and Kish which were just then coming to light, Second, a considerable number of Pahlaviworks and the palace acquired a . describing Sasanian court ceremonial survived There is no doubt that a very large building into the later Middle Ages and were used by once stood on this platform, and it may well have Muslim chroniclers. The Kitab al-taj of Jahiz (d. been the palace of Khusraw mentioned by the 869), for example, seems to have incorporated medieval geographers. Butbefore using this draw- much material from the Gahnama, a notitia digni- ing to discuss the nature of Sasanian gate com- tatum of the Sasanians which listed according to plexes, the typical Sasanian arrangement of rank all the dignitaries of the Persian monar- domed hall fronted by an iwan, or the basilical chy.20 As vital as such sources are for an under- hall in Sasanian architecture, we should dwell for standing of internal politics in the royal court, a moment on its pedigree. Bell informs us that in they provide virtually no direct information about producing her survey she was sometimes obliged an architectural background. to make analogies with the better-preserved pal- The archaeological evidence for continuity of ace at Ukhaidir in Iraq to fill in the missing form and function is no less equivocal. The prob- parts,18 of which there were many. I suspect this is lem is best illustrated by considering briefly the why Khusraw's building has such a strong Abassid setting for the audience. Very few Umayyad palac- flavor. Put less delicately, it seems to me a fine es, first of all, preserve locales that can be identi- example of how Sasanian architecture can be fied with reasonable certainty as throne rooms. influenced by early Islam. Two of these are Mshatta and Khirbat al-Mafjar. AtMshatta2 l the throne complex lay at the back of The assumption that architectural design in any the walled enclosure directly opposite the en- period is somehow influenced by that of the trance gate, and consisted of a triconch preceded preceding one is not only reasonable but an by a long hall open at the front that was divided underlying principle of architectural history. Due into a broad central nave flanked by side aisles. At largely to the dearth of reliable archaeological Khirbat al-Mafjar 22 the audience most likely took data at the Sasanian end it has not been possible place, as Ettinghausen once demonstrated,23 in a to define systematically the nature and extent of complex that included a pillared hall with a this relationship between the palaces of the Khus- broad central aisle that led from a gate structure raws and those of their Muslim successors. Stud- to an apsidal room at the back. The ensemble was ies have tended to focus on isolated features such richly decorated with mosaic and stucco that as the four-iwan plan and the familiar combina- incorporated an elaborate program of images tion of iwan and domed hall. taken from Sasanian royal sources. Most striking Two classes of evidence have fostered the wide- are the stucco figure of a prince in Persian dress spread notion that there was a continuity in added to the gatehouse facade at a later time, and palace design in a more comprehensive sense, the stone chain and headdress which hung from but they are largely circumstantial and of limited the semidome, presumably above the throne. significance. The first is an extensive body of Two points can be made here, the first being symbols and imagery originally associated with that neither the triconch nor the pillared hall is Sasanian kingship which survived in all media known in Sasanian palace architecture and in- into the Umayyad period and later. Grabar, in his deed would seem to be quite uncharacteristic. doctoral thesis of 1955 and in a number of later The second is that while the Umayyad audience publications, 9 has dealtin great detailwith Umayy- could apparently take place in any number of ad ceremonial as it is described in the Arab architectural settings, the Sasanian audience was sources, relating it to the material remains as connected primarily, if not exclusively, with the these have become available. He has shown how iwan hall, with or without a domed chamber in the Umayyad rulers were able to create for them- back. This is certainly the impression one gets selves an ambiance of princely splendor that was from the Muslim sources which deal specifically drawn in large measure from the defunct Persian with the Arch of Chosroes. But the Sasanian court. He did not, however, press the issue of monuments themselves insofar as we know them continuity of its architectural setting, noting that give the same impression. the desert castles of Syria, Jordan, and Pales- The so-called Taq-i Girra, which probably dates tine-virtually all thatremains ofUmayyad prince- to the Middle Sasanian period, seems to repro- ly architecture-derived from local Roman and duce the form of an iwan hall, and cuttings in the Byzantine traditions. floor and at the back suggest that it held a statue, 60 LIONEL BIER most likely a royal one.2 4 The rock-cut iwans at derives from a vocabulary of decorative motifs Taq-i Bustan, richly decorated in relief with royal clearly originating in Sasanian stucco. The artic- imagery, may actually have been provided with a ulation of the court wall of the qar, a kind of throne. 25 In Qala-i Dukhtar, the royal audience entrance building, with tiers of niches framing certainly took place in the great iwan hall at the the iwan arches, makes, on a miniature scale, an center of the building. Huff, noting the window emphatic allusion to the Taq-i Kisra at Ctesiphon, opening high at the back, and a fragmentary the great palace of the Sasanian kings.3 ' stone basin discovered in the middle terrace, There have been attempts to establish a second compared the arrangement with seventeenth- type of Sasanian audience complex based on century pavilions in which accommodat- what are in fact strong similarities between the ed the Safavid audience and had windows in the building at Damghan at the core of the Umayyad upper story from which courtiers could view the daral-imaraatKufa. 32 But the two major buildings official activity taking place below.2 6 normally pressed into service to form a class- In a detailed analysis of Mshatta, Hillenbrand Sarvistan and Qasr-i Shirin-are of dubious val- plays down the importance of the forms of the ue. Since Sarvistan can no longer be attributed to individual halls as indicators of Sasanian influ- the Sasanians, 33 and since the Imaret-i Khusraw is ence, stressing instead their arrangementwith an a fantasy based partly on Damghan itself, the open court along a single axis: "Functionally, arrangement at Damghan must remain an anom- there is very little to choose between the Partho- aly, one whose precise function is unclear. Sasanian formula of an iwan preceding a domed There is ample evidence that the Abbasid ca- chamber and the classically inspired formula of a liphs followed the Umayyads in incorporating basilical hall preceding a triconch audience cham- Sasanian practices into their ceremonial,3 4 but ber."27 He continues his general argument for their palaces, insofar as we know them from Sasanian influence in late Umayyad palace archi- Samarra and isolated monuments like Ukhaidir, tecture by pointing to the vaulting in this official had fewer affinities with Sasanian architecture area at Mshatta, suggesting first that its pitched than might be expected. They are characterized brick construction was inspired by Sasanian ar- by their sprawling plans that contained a great chitecture-most likely the palace at Ctesiphon, number of units, and consisted most typically of where the brick rings also incline towards the rear courts in series connected by gate structures wall-and second, that the very use of brick vault- around which were grouped numerous bayts of ing in the audience complex of a stone building fairly uniform format.3 5 may have been intended as a reference to the The Abbasid gate complexes were architectur- Taq-i Kisrawhere brickwas also used "for the area ally significant and had ceremonial importance.? most closely associated with the sovereign. "2 The But whether these or the Umayyad palace gate- fact remains, however, that while the great vault ways before them owed anything to the Sasanians of Ctesiphon is indeed built of brick, the rest of is a moot question, as little Sasanian gate architec- the building is also, and tunnel vaults of pitched ture survives. The unit is known only in the very brick laid vertically or in inclined rings are com- early Qala-i Dukhtar where the layout and built- mon enough in .29 Thus, in features suggest a reception area rather than a while it is true that Mshatta has a strong Iranian place of appearances. 3 7 flavor, the nature and extent of Sasanian influ- Sasanian influence becomes a real factor only ence is difficult to define. It seems to have consist- with the cruciform grouping of rooms which ed of little more than an axial disposition of the were clearly the focus of these complexes. There halls and court at the official center of the palace are two variants. The first, found in all the major and the deployment of Sasanian royal symbols in palaces of Samarra, consisted of four axial iwans the carved ornament. fronting a domed room.38 In the second, repre- An intriguing example of Umayyad palace ar- sented at Samarra only by the "resthouse" behind chitecture of some relevance here is the complex the mihrab of the Abu Dulaf , four iwans at the northern edge of the Amman citadel, open on a central court.3 9 The first type is known which seems to have been built and decorated in also from the daral-imara of Abu Muslim at the Sasanian mode.3 0 Constructed in the local cut- and probably formed the nucleus of the great stone technique, its nucleus consisted of a domed palace of al-Mansur at .4 chamber fronted by an iwan hall that opened on To re-create the ceremonial of an earlier an inner court. Its unmistakably Persian aspect dynasty one needs a reason for doing so and some THE SASANIAN PALACES AND THEIR INFLUENCE IN EARLY ISLAM 61

genuine text to serve as a guide. Something of the shown that the palaces there were occupied physical ambiance can be reproduced by copying into the early Islamic period, butwe do not know a courtly style of stucco decoration or metalwork how the buildings were used or how far their from examples that in the early Islamic period physical integrity was appreciated and respect- must have survived in ample quantities. Continu- ed.4 ' At Takht-i Sulayman we have in the Ilkhanid ity in ceremonial practice also implies a parallel period a rare case of builders incorporating Sasa- re-creation of architectural features to provide a nian walls which then determined the plan of the proper framework. new palace.42 It is very likely that the audience ensembles of As a result of recent survey work at Samarra, the both Abbasid and late Umayyad palaces derived remains of a large Sasanian palace have been from Sasanian models. Whatever symbolic value identified immediately adjacent to the Qasr al- such appropriations might have had for the early Jafari of al-Mutawakkil. 4 3 The Sasanian building Muslims, these ensembles were eminently suited was renovated when the Abbasid palace was con- to an audience ceremony that began with a reve- structed in 859-62, at which time a substantial lation in which the ruler was stationary. water tank with supply channels and drains was It might be useful at this point to consider the built into it. Certain features, notably a series of question of what early Muslim builders and their courtyards and public rooms, have been tenta- princely patrons could have known about the tively located in the unexcavated debris, and palace architecture of the Sasanian kings. In there was apparently a hunting enclosure nearby Western tradition, concepts of architectural plan- which was used into the Abbasid period. Further ning and details of construction are often trans- work there may provide some insight into the mitted by a process involving the close observa- caliph's attitude towards the buildings of a Sasa- tion of existing monuments and their description nian predecessor. in architectural treatises. This process is exempli- The rulers of certain Iranian dynasties of the fied by Vitruvius, that Roman architect of the first early Islamic period must have had a special century B.C. who traveled about examining earli- interest in Sasanian palace architecture. The er monuments of architecture in order to estab- Muslim Buyids, for example, traced their lineage lish principles for practicing architects and build- back to the Sasanian kings, and a few of them are ers of his own day. known to have displayed an active interest in the This antiquarian, indeed forensic, approach to ancient monarchy.4 'Adud al-Dawla, who rebuilt architecture is nowhere in evidence in the early the Sasanian capital of Gur, renaming it Firuza- Islamic period. We do have references in geo- bad, appears as a Sasanian ruler on coins minted graphical and historical works to dozens of palac- in Fars that bear the Persian title shahanshah. He es, princely residences, hunting lodges, and gar- proudly recorded at a visit he made to den pavilions. Their authers refer to wonders of the site in the company of a mobad from Kazarun construction and decoration such as columns in who read to him an inscription in Pahlavi. The the shape of women and blocks of stone so finely founder of the line from which the Buyids worked that the joins were invisible. Their main emerged is said to have dreamed of conquering purpose, aside from marking a conspicuous fea- Iraq, rebuilding the palace at Ctesiphon, and ture of a locale, was to impress the reader with reestablishing an Iranian state based on the an- certain qualities of the original occupant. But cient Zoroastrian religion. Unfortunately, we have they contain no information that would enable a little Buyid architecture of any kind, and none of builder to understand these remarkable monu- the palaces thatwere described by contemporary ments in architectural terms. authors have survived.45 Even imagining a prince with archaeological Finally, I would like to return to the great inclinations, it is difficult to say, given the paucity palace in Ctesiphon which was probably erected of available data about the post-Sasanian histo- by Khusraw II in the sixth century. When the Arab ries of the Sasanian palaces, what information commander entered the Sasanian capitol in 637, remained to be gathered. Some were already in he led the Friday prayers in the throne hall and, ruins when Yazdigird fled the capital for the from that moment, the building assumed great Iranian plateau. Dastagird and Qasr-i Shirin, for symbolic significance to the Muslims. This is example, had been totally demolished by Herac- perhaps most dramatically expressed in the lius in the sixth century. often cited passage in al-Khatib al-Baghdadi's Excavations at Firuzabad and Bishapur have introduction to his history of Baghdad which 62 LIONEL BIER describes al-Mansur's demolition of the Iwan-i that the influence of the Sasanian palaces on Kisra and the reuse of its bricks for his own early Islam was largely in the realm of poetry and palace.46 He relates how al-Mansur proceeded metaphor. despite a council of non-Arab advisers who ar- There is no doubt that early Muslim rulers gued that the palace was a monument to the Arab looked to their Sasanian predecessors for means victory over the Persian kings, but how the caliph by which to express a concept of kingship in desisted only when the undertaking proved too architectural as well as ceremonial terms. But the vast. Al-Tabari offers a variant in which an adviser resulting adaptations were usually so subtle and now recommended pushing on at all costs lest the complete that they defy attempts to isolate the caliph's inability to destroy the palace damage his various components. There is no evidence that prestige in the eyes of his Persian subjects. 47 early Muslim princes sought to imitate the Sasani- Whether or not such anecdotes reflect histori- an palace in a comprehensive sense and it is cal reality, they are interesting because they illus- doubtful that there was readily available suffi- trate what seems to be the real significance the cient archaeological information to do so. When monuments of the Persian kings had for their Sasanian influence is evident at all, it is invariably Muslim successors. When we consider these ac- seen in the official portions, more specifically in counts alongside the quasi-historical traditions the throne-room ensemble which must have em- and romances which later grew up around this bodied for writers and builders alike the essence and other Sasanian monuments like Taq-i Bustan of Sasanian imperium. and Takht-i Sulayman in Iran,4 it becomes clear THE SASANIAN PALACES AND THEIR INFLUENCE IN EARLY ISLAM 63

Notes Court," Ph.D. diss., Princeton, 1955; "Notes sur les ceremonies umayyades," Studies in Memory of Gas- 1. In Arthur Upham Pope, ed., A Survey of PersianArt ton Wiet, ed. Myriam Rosen-Ayalon (Jerusalem, (London, 1938) 1:493-578. 1977), 51 ff.

2. Erich Schmidt, Excavationsat TepeHissar,Damghan 20. See Arthur Christensen, L'Iran sous les Sassanides (Philadelphia, 1937), 327 ff. and fig. 170. (Copenhagen, 1944), 62 f.

3. Pope, Survey, 1: fig. 166 (drawn by Oscar Reuther). 21. See K. A. C. Creswell, Early Muslim Architecture, 2 vols. (Oxford, 1932-40), 1:578 ff. and plan. 4. Pope, Survey, l:fig. 167. 22. R. W. Hamilton, Khirbat alMafjar:AnArabianMan- 5. Reproduced in P. R. S. Moorey, Kish Excavations sion in theJordan Valley (Oxford, 1959). 1922-33 (Oxford, 1978), fig. J. 23. Richard Ettinghausen, From Byzantium to Sasanian 6. Moorey, Kish Excavations, 122 ff. Iranand the Islamic World (Leiden, 1972), ch. 3. But see R. W. Hamilton, "Khirbat al-Mafjar: The Bath 7. Georges Salles and Roman Ghirshman, Bichdpour, Hall Reconsidered," Levant 10 (1978): 126 ff., who vol. 2: Lesmosdiquessassanides(Paris, 1956), passim. sees this complex as Walid's majlis al-lahu and See plan II. denies that ornament was consciously used to assert legitimacy. See also his Walid and HisFriends 8. My own compass reading taken in 1976 showed a (Oxford, 1988). variance of about two degrees. 24. Hubertus von Gall, "Entwicklung und Gestalt des 9. Encyclopaedia Iranica, s.v. "BiSapir," 4:3, 287-89. Thrones im vorislamischen Iran," Archdologische Mitteilungen aus Iran, n.F. 4 (1971): 221 f. 10. Salles and Ghirshman, Bichdpour, 2:pl. I. 25. Von Gall, "Entwicklung und Gestalt," 221. 11. Klaus Schippmann, Die iranischenFeuerheiligtiimer (Berlin, 1971), passim. 26. Dietrich Huff, "Qal'a-ye Dukhtar bei Firuzabad," Archdologische Mitteilungen aus Iran, n.F. 4 (1971): 12. Massoud Azarnoush, 'Tire Temple and Anahita 164 ff. Temple: A Discussion of Some Iranian Places of Worship," Mesopotamia 22 (1987): 393 ff. 27. Robert Hillenbrand, "Islamic Art at the Cross- roads: East versus West at Mshatta," in Essays on 13. See Ali Akbar Sarfaraz, "Anahita, Ma abad-e Islamic Art and Architecture in Honor of Katharina Bozorg-e Bi'tapfr," in Proceedingsof the IIIrdAnnu- Otto-Dorn, ed. Abbas Daneshvari (Malibu, Calif., al Symposium on ArchaeologicalResearch in Iran, 2nd 1981), 63-86, esp. 71 ff. to 7th November, 1974 (Teheran, 1975), Persian section, 99. 28. Hillenbrand, "Islamic Art at the Crossroads," 72.

14. Hubertus von Gall, "Die Mosaiken von Bishapur," 29. For pitched brick construction with vertical and Archdologische Mitteilungen aus Iran, n.F. 4 (1971): inclined rings, see John Ward-Perkins, "Notes on 221 f. the Structure and Building Methods of Early Byz- antine Architecture," in The Great Palace of the 15. Pope, Survey, :plan, fig. 153 with reconstruction, Byzantine Emperors, Second Report, ed. David Talbot fig. 154. Rice (Edinburgh, 1958), 580 and passim.

16. J. de Morgan, Mission scientifique en Perse, vol. 4 30. See Alastair Northedge, "Survey of the Terrace (Paris, 1896): pls. 40, 42, and 46. Area atAmman Citadel," Levant 12 (1980): 150 ff.; 'The Qasr of Amman," Art andArchaeology Research 17. Gertrude Bell, Palaceand Mosque at Uhhaidir(Ox- Papers15 (1979): 26 f. See also the brief discussion ford, 1914), 44-51 and pls. 53, 54. byJ. W. Allan in Muqarnas8 (1991): 13 f.

18. Bell, Palaceand Mosque, 44-51. 31. Northedge, 'The Qasr of Amman," 26.

19. Oleg Grabar, "Ceremonial and Art at the Umayyad 32. Forexample, Oleg Grabar, "Al-Mushatta, Baghdad, 64 LIONEL BIER

and Wasit," in The World ofslam, ed.James Kritzeck 42. Rudolph Naumann, ArchiiologischeAnzeiger (1965), and R. Bayly Winder (London, 1959), 104. 697 ff.

33. For a date of construction in the ninth century, see 43. Alastair Northedge et al., 'Survey and Excavation Lionel Bier, Sarvistan: A Study in Early Iranian at Samarra, 1989," Iraq52 (1990): 132 ff. Architecture (University Park, Penn., 1986), passim. 44. For Buyid interest in the Sasanians, see C. E. 34. Dominique Sourdel, "Questions de crmonial Bosworth, The Heritage of Rulership in Early abbaside," Revue des Etudes Islamiques 38 (1960): Islamic Iran and the Search for Dynastic Connec- 121 ff. tions with the Past," Iran 11 (1973): 51 ff. and H. Busse, "Iran under the Buyids, in Cambridge 35. Ernst Herzfeld, Geschichte der Stadt Samarra (Ham- 4:273 ff. Also, Richard Frye, "The burg, 1948), passim. New Persian Renaissance in Western Iran," in ArabicandIslamicStudies inHonourofHamilton A. R 36. See Grabar, Ceremonial and Art, 125 ff. Gibb (Leiden, 1965).

37. Dietrich Huff, "Ausgrabungen aufQalPa-ye Dukhtar 45. Muqaddasi, for example, reported that 'Adud al- bei Firuzabad 1976," Archdologische Mitteilungen Dawla built a palace with 360 rooms, each decorat- aus Iran, n.F. 11 (1978): 117 ff. and fig. 1. ed in a different style, in the vicinity of . See Donald Whitcomb, Before the Roses and Nightingales: 38. See Yasser Tabbaa's discussion of the four-iwan Excavations at Qasr-iabu Nasr, Old Shiraz (New York, plan in this volume. 1985), 140 ff., for the topographic problems.

39. Most recently, Alastair Northedge, Muqarnas 8 46. Jacob Lassner, The Topography of Baghdad in the (1991): 89 and fig. 10. Early Middle Ages (Detroit, 1970), 128.

40. For a recent summary of attempts to reconstruct 47. Lassner, Topography of Baghdad, 128. the plan of the palace at Baghdad, seeJ. W. Allan, "New Additions to the New Edition," Muqarnas 8 48. See, for example, Gerd Gropp, "Neupersische (1991): 17 ff. Uberlieferungen vom Heiligtum auf dem Taxt-e Soleiman," Archaologische Mitteilungen aus Iran," 41. For Qala-i Dukhtar, see Dietrich Huff, Archdolo- n.F. 10 (1972): 243 ff.; Priscilla Soucek, "Farhad gische Mitteilungen aus Iran, n.F. 9 (1976): 173; 11 and Taq-i Bustan: The Growth of a Legend," in (1978): 140. Occupation of the palace of Bishapur Studies in Art and Literatureof the NearEast:In Honor during the early Islamic period is attested mostly of RichardEttinghausen,ed. Peter Chelkowski (New by decorative stucco and coins. See Salles and York, 1974). Ghirshman, Bichdpour, 2:149-99.