P. Pfiilzner courtyard paved with baked bricks, through which a large hall, probably a reception room, was accessible via a recessed door. On the courtyard side the wall or this hall is buttressed. This arrangement is strongly reminiscent of the public wings in the afore-mentioned EJZ 3b palaces, the Bcvdar central palace, and also Palace F at Khuera. Inside the large room is a plariorrn, leading the excavators to suggest that the room served both as a banquet and throne room.220 Behind the reception room there is another large room, which might alternatively be regarded as a throne-room.221 The Eastern Palace is not vet completely excavated thus no final conclusion on its overall functional organisation can be drawn. Nevertheless, a typical scheme of EJZ 3 palaces can be deduced from the three examples presented. 5.4,3 Period EJZ 4 Palatial architecture is characterised by a discontinuity between Periods EJZ 3 and EJZ 4. At Khuera, Palace F is transformed into a workshop area for pottery production in Level la-b, which corresponds to Period Khuera I E (= EJZ 4).222 The palace had been gradually abandoned during the earlier Level 2a and the workshops of Phase 1 installed in its ruins. Ac Beydar the palace in Area F was totally abandoned in Period EJZ 4.223 This hints at a general discontinuity in political order between the two periods. A different picture is visible at Mozan/ Urkesh. Here, the so-called "Tupkish Palace" (or Palace AP) was constructed in Period EJZ 4. Its two main Phases, 3 and 2, date to this period/2,1 The huge palace, still incompletely excavated, is divided into two separate wings located on different terraces: the formal wing to the E, and the service wing to the W (Fig. 47).225 The formal wing is arranged around a huge central courtyard paved with large flagstones (H3).226 It is remarkably different from the baked brick pavement of the palace at Beydar, but has a parallel in the courtyard of Khuera's Palace F (see above).22. The courtyard at Mozan is bordered on the S side by a unit of small regularly arranged rooms (H4-6,11-3). A reception room and throne room within the formal wing, typical of EJZ palaces (see above), has not vet be exposed. The service wing consists of two parallel room units with a similar layout. Each unit is arranged around a long central room (A5/ C5) with an iwanAiht annex to the S (A2/ C2).228 This www-type of room is already attested in Period EJZ 3b in the palace at Beydar (see above), but at Mozan it it does not have pillars at the front. The central room of this tripartite unit has a lateral wing on both sides consisting of two or thtee smaller rooms. At the N end each unit is closed by a broad room (B /D) separating the tripartite unit from a brick paved courtyard (F) to the N, ftom where both symmetrical units were accessible. A peculiar feature is the existence of a tiny chamber (1.8 x 1.3m), called a "closet" or a "vault" by the excavators and situated in one corner of each broad room in the tripartite unit.229 Both closets are located perfectly opposite to each other within the two broad rooms, upholding the overall symmetry of the tripartite units. They are accessible through an "iwan"-like annex to the broad room. The closets, which have thick walls and a recessed entry, are interpreted as having served for the storage of special goods. The two units of the service wing contain an oven and a hearth,230 a toilet,231 a basin (possibly a scribal installation),232 and a subterranean drain.233 The floor deposits yielded many seal impressions,23'' thus an economic as well as a residential function can be assumed. When compared with other EJZ palaces, the tripartite units are remarkably reminiscent of the residential wing on the upper terrace of Palace F at Khuera, which has a similar tripartite layout (see above). The two separate units in the Mozan palace were perhaps used as living and storage areas for different social units within the palace. Tire integration of two parallel tripartite units into one sector of the palace is a very peculiar arrangement unparalleled at othet EJZ palaces. Furthermore, the "closets", the /^^-arrangement, and the symmetrical layout are unique characteristics of the Tupkish palace. They provide this building with a very distinctive character compared to other EJZ architecture. 2:0 Pruss & Schmitt 2009: 19 f„ Plan p. 26. 221 Pruss 2008: 16, 19. 222 Orthmann & Pruss 1995: 124 £ 223 Lebeau 2003b: 26. 224 Buccellati & Kelly-Buceellati 2002a: Fig. 2. 225 Buccellati & Kelly-Buceellati 1995a; 1995b; 1995-96; 1996a; 1996b; 2000; 2001; 2002a; 2004; Buccellati 1998; 2005. 220 Buccellati & Keliy-Buccellati 2004: 14-19, Fig. 2-4. 22~ Compare: Orthmann & Pruss 1995: 122. 228 Buccellati & Kellv-Buccellati 2000: 142 t. (the excavators introduced the term "swan" tor the S part C.2 or room C5). 225 Buccellati & Kelly-Buecellati 1995a: 391, Fig. 4; 1995-96: 4 f., 28: Fig. 3, Photo 3. ""Buccellati 1998: 21, 27, Fig. 5. 231 Buccellati & Kelly-Buceellati 2000: 142. 232 Buccellati & Kelly-Bucccllad 2000: I4I-I46. 233 Buccellati 2005: 19-21. 234 Buccellati & Kelly-Buceellati 1995-1996: 6 it.; 1995b: 65 ff. Architecture Fig. 47: Mozan, Tupkish Palace, Phase 2-3, EJZ 4 (Buccellati 2005: Fig. 3; modified by G. Buccellati) (with kind permission of G. Buccellati). At Leilan the so-called "Akkadian Palace" belonging to Period EjZ 4 is recorded on the NW Acropolis (Fig. 48).235 It is attributed by the excavators to Leilan Period lib, corresponding to the Akkadian period (EJZ 4). The building, as yet incompletely excavated, is fortified and bordered by a glacis. Inside the 6.6m wide defensive walls are several rooms containing a large number of ovens and interpreted as the grain processing facilities or an administrative institution. Two other rooms were identified as a granary and a "tablet room". Besides this, no other palatial features such as a courtyard, reception or throne room, or other typical architectural units have been recorded yet. Thus, no clear functional or architectural characterisation of the building is currently possible. There is a later occupation of the palace area during Leilan Period lie (= end of EJZ 4 or EJZ 5), dated by C14 samples to around 2140 BC, but it is not clear whether the four remaining rooms maintained palatial functions. The most famous, and most disputed, palace of Period EJZ 4 is the Palace of Naram Sin at Brak, excavated by Mallowan in 1937/38236 and investigated further in later years.237 The building is clearly dated to the Akkadian period (EJZ 4) on the basis of inscribed bricks with the name of the Akkadian king. The huge building measures 111 x 93m in size, two thirds of which have been excavated (Fig. 49). Only the foundations of the building survived, making architectural reconstruction difficult. The regular architectural layout is structured around a huge square central courtyard, 41 x 41m in size, and three smaller courtyards in the N part of the building, arranged parallel to each other with regular dimensions of 18 x 18m, 18 x 15m and 15 x 15m. All courtyards are surrounded by a single or double row of oblong rooms, all with a standardised width of 2.2m, and probably used for storage. No larger, typical palatial rooms were documented. Based on the enormous rhickness of the 10m wide outer wall and on the interior arrangement of rooms, it has been proposed rhat the building was a "fortified storehouse". -Tli is identification is supported by the discovery of burnt grain in three of the storerooms23'' and cuneiform tablets mentioning people and commodities such as grain.240 Thus, the building probably served a purely administrative and economic purpose. This contrasts with other EJZ palaces, with the possible exception of the Akkadian Palace at Leilan.241 There is, despite the differences in dimension, a striking similarity of the Narim Sin Palace with the layout of the Pusham House at Mozan (see above). This might reflect similarities in the economic function of both buildings, albeit on a different socio-political scale. 235 All information given here is taken from the Leilan webpage: http://leilan.yale.eciu/abour/dig_sites/acropolis_northvvesr index.html; see also: de Lillis Forrest et. al 2004; de Lillis Forrest, Milano & Mori 2007; Weiss et al. in prep, (the palace is also referred to as "Dudu Palace", but this provisional designation has been dropped: H, Weiss, personal comm.). ** Mallowan 1947: 63-68, PI. LIX-LXI. 237 Oates & Oates 2001a: 19-21, Fig. 15-19. iw Oates & Oates 2001a: 19: Oates 2007: 171. 239 Mallowan 1947: 63 f. 2<0 Mallowan 1947: 66: Gadd 1940: 60 f. 2,1 The latter needs to be more fully excavated in order to verify this assumption. P. Pfalzner Leilan 2008 Geneva! Flan Akkadian Phase 0 5 10 ——i Meters Fig. 48: Leilan, Acropolis Northwest, Akkadian Palace, Leilan Period lib, EJZ 4 (Leilan webpage; and: Weiss et al. in prep.) (with kind permission of H. Weiss). Fig. 49: Brak, Area CH, Naram Sin Palace. Brakperiod M, EJZ4 (Mallowan 194 : PI. LX). 75 Architecture r,.A.i PeriodEJZ5 Evidence of Period EJZ 5 palaces is sparse. In view of the historically attested existence of a local i li dynasty at Urkcsh, there should theoretically have been a palace in Mozan at this time.212 However, it is stri to observe that the Tupkish Palace was abandoned at the end of Period EJZ 4.2"15 The area of the former palact transformed during Period EJZ 5Z'~' into a domestic quarter with scattered housing and open areas for d< activities. Probably, a palace was constructed at an alternative location within the city, although this still nc. be round. It is interesting to note that a royal seal, probably of a local ruler, was found on impressions in the 5 Pusham House. This indicates that the economic activities of this building were at least partly interrelate: the royal household of Urkesh at this time.2" At Brak, the Naram Sin Palace was reused during Period EJZ 5. As Mallowan discovered, the wall-Akkadian palace were cur down ro floor level and new walls erected, following, with the exception ol . 32, the outline of the earlier building. The floors were raised by 3m. The rebuilding of the palace was attri'r by Mallowan to the Ur Ill-period (= EJZ 5). This correlation was based on a tablet found in Court 2. debris ot the later building, and probably bearing the name or Ur-Nammu,-"iG as weii as a seal inscript Ur-Nammu, which came ro light in a rubbish dump.2'1 The existence of Ur-Nammu's name in these in-tions has been ruled out by a later re-study of the texts.248 Thus, an Ur-III date for the later palace is uncc but. a post-Akkadian, Brak period N (= EJZ 5) date of the Naram Sin successor building can still be str assumed on purely stratigraphic grounds.u^ Thus, it still can be counted as the onlv known palace of ! : EJZ 5.250 5.5. Temples and Ceremonial Architecture 5.5.1 Period EJZ 0 and EJZ 1 There is not yet any evidence for temple architecture in Periods EJZ 0 and 1. 5.5.2 Period EJZ 2 An early Syrian JZ temple has been documented at Raqa'i, Level 3 (EJZ 2).2,1 It is a free-standing shri:_ 4.5m in size, erected within an open courtyard surrounded by an enclosure wall (Fig. 50).2,2 The "cella" is a : axis room accessed through a door framed bv buttresses and locared close to the N end of the room. In rhe int there is a stepped altar podium on the N wall, flanked by a second, lower podium. In the middle of the roi 1 floor level, are the remains of a hearth. Two small side rooms, accessed separately, are attached to the N -the "cella". The structure has been defined as a temple, primarily because of its conspicuous divergence in from the other houses at Raqa'i and the discovery of an altar.2'5 This interpretation has been challenged di the similarity of the rooms and particularly of the altar podium to the houses at Khuera with their house alta Consequently, the so-called "temple" at Raqa i could be interpreted as a Khuera-type house with an ancestor or as a small shrine for family ancestors. 2,2 As generally accepted, this is the time of the Human kings Shatar-mat und Atal-shen (Buccellati & Kelly-Buccellati 20 f. Fig. 19; 2002a: 113-115; 2005; 42-44, Tb. 2.). 243 Buccellati & Kelly-Buccellati 2002a: 106-109, Fig. 2-3. 244 Buccellati 8c Kelly-Buccellati 2002a: 113 f.; 2004: 15, Fig. 8d.; 2005: 29; Buccellati 2005: 23; Pfalzner Sc Dohi Pfalzner, in press. 2,5 It is the seal impression of King Rimush, not to be equated with Rimush ot Akkade, but probabl) a local ruler ot Urkesi ing the Ur III period (Volk 2004: 95-98, Abb. 4; Pfälzner & Dohmann-Pfälzner, in press; Pfälzner, in press). 246 Mallowan 1947: 69; Gadd 1940: 61 (last paragraph); unfortunately the reading of the king's name is questionable an tablet probabl)' a writing exercise. 247 Mallowan 1947: 69. 248 Eidem, Finkel (sr Bonechi 2001: 109 (No. 23), 111 (No. 40). 249 Oates & Oates 2001a: 15-19. 150 The palace lias recently been connected with the reign ot the local Human king [alpuS-atali, dated to the post-Akk. time (including the Ur III period); see: Gates & Gates 2001a: 1 5; 2001 b: 393 f., Fig. 376. r,! Level 3 is dated by Quenet, this volume to the final FjZ 2 period. 252 Curvers & Schwartz 1990: 12-13. Fig. S, 12; Schwartz 2000: 167-170. 03 Curvers & Schwartz 1990: 12; Schwartz 2000: 170-1~. 254 Pfalzner 2001: 175, 309. P. Pfihner the mound. Two rooms and tananir could be identified, at least demonstrating that there was domestic occupation at the site during Period EJZ 5.2W In summary, the few substantial examples of EJZ 5 domestic architecture at Mozan and Brak represent large, well-built and prosperous houses. This contradicts the general assumption of impoverished urban living conditions during this last phase of the 3"1 millennium. It rather hints at the existence of wealthy households involved in supra-regional activities during this time.204 5=4. Palaces In line with the genera! development of political structures in the EJZ period, no palaces are attested in periods EJZ 0, EJZ ] or for most of EJZ 2. Tie earliest palatial building attested in the region dates to the end of Period EJZ 2. 5-4,1 Period EJZ 2 The oldest palatial building dating to the EJZ period is the so-called "Leilan 11 Id Palace", situated on the city's Acropolis (Fig. 42). Tie part of the building excavated corresponds to Levels 17-16, dating to the end or Period EJZ 2.205 The wails are up to lm wide and it has two groups of storage rooms, one comprising nine rooms, and the other three. More than SO ED Ilia style seal impressions were found. These were associated with the building itself and an ash layer on top of it. Most of the storage rooms are very small, the largest one (Room 1) measuring 2.6 x 2,3m. In view of this and the fact that sealings arc also widely present in EJZ period domestic contexts,20" it is possible that this complex of rooms was not actually palatial in nature. It could equally have been the storage area of a non-palatial communal, religious or private context. Thus, EJZ 2 palatial architecture remains obscure. 5.4.2 Period EJZ 3 Aside from the ambiguous Leilan (Period IHd) evidence, JZ palatial architecture is first fully attested in Period EJZ 3. Palace F at Khueta has been documented for three phases: Building Levels 2b and 2a date to period "Khuera ID late", whilst Building Level 3 dates to period "Khuera ID early".20" Both of these local phases correlate with Period EJZ 3b. The largest exposures pertain to Level 2, whilst Level 3 possibly represents an earlier phase of the building, only reached by soundings in a few places.203 The palace is not centrally located, but is positioned with its back attached to the inner city wall. In the main building phase, Level 2b, the palace is strictly rectangular in outline, although it is terraced (Fig. 43). The central residential area is located in the upper terrace Fig. 42: Leilan, Acropolis, Levels 17-16, so-called "Leilan Hid palace", EJZ 2 (Calderonc & Weiss 2003: Fig. 3). 203 Tunea&Miftah200": 13F,Fig. 2.18a. See: Pfalzner 2010: 4-10; in press. 205 Weiss 1990b: 209-213; Calderone & Weiss 2003: 194 197. 206 See: Pfalzner 2001:232-239. 207 Orthmann. Sc Pruss 1995; 121 f., 139 f. 208 Ibid. 122-124, PL 17-20, Beilage 15: Pruss 1998a; 2 1 b; Orthmann 1990b: 25, Fig. 19. Architecture / Ail ¥ GftrtJfds/BkciaJrJia.-. j CM***,** £ Fig. 53: Mozan, Area B, Temple terrace. Stage II, EJZ 3. At Kashkashok III, Area A, there is one building that may have functioned as a temple. It is a freestanding, single-roomed structure (Room 13) measuring 6 x 8m with thick outer walls and a bent axis entrance (Fig. 51).:" The interior has an altar on the W side as well as buttresses and benches on both the N and the W sides, The building belongs to Level AIV, which dates roughly to the end of Period EJZ 2 or the beginning of EJZ 3.2'f" The closest parallel is to the so-called "temple" at Raqa'i, although this too cannot be firmly identified as a cultic building (see above . At Brak, a temple was excavated in Area HS4 (Fig. 52). This was initially constructed in Level 5 and rebuilt in Level 4.25' Both levels date to Period Brak K (EJZ 2). The structure is a single-roomed-shrine of 8 x 4.5m although the location of the entrance is unclear. Inside there are benches along one long and one small wall and a free-standing altar along the central axis. In front of this, at floot level, is a rectangular hearth. Despite the correspondence in date, this structure is clearly distinct from the Raqa'i "temple" with regard to interior organisation: at Raqa'i. the altar is not detached from the back wall, nor does it have benches. These dissimilarities hint at a difference in function. Nevertheless, due to the lack of distinctive religious objects inside the cella of the Brak temple, it~ identification as a temple remains inconclusive. Indeed, the same set of installations, with the exception of a freestanding altar, are also charactetistic of EJZ houses.258 The huge temple terrace at Mozan/Urkesh, located on the high tell in the city's centre, was built during Period EJZ 2 (Fig. 53). Two floor levels were identified (Phases BS 6a and 6b) both dating to this period. These were connected to the mud-brick ramp substructure leading up to the high mud-brick terrace.23' These structures define Stage I of the temple oval. Thus, the temple at Mozan is the oldest known monumental temple building in the Syrian Jezireh. The exact layout of the temple in Period EJZ 2 is not yet known. The existence of a ramp and high terrace (see below) is clear. The ramp did not vet have its broad horizontal stone revetments. Also, there is no evidence for the existence of the stone-built, oval temenos wall in this early period. Nor is the temple building, which stood on top of the terrace, yet attested with a floor level of Phase EJZ 2, although it can be hypothetically assumed that the EJZ 3 cella on the terrace (see below) already existed in this period. 5.5.3 Period EJZ 3 Period EJZ 3 is characterised by the existence of large temple precincts. The temple oval at Mozan/ Urkesh is one of the most prominent examples. The huge mud-brick terrace, designated Stage II of the temple oval, evidently existed at this time.260 It is a monumental construction of mud-bricks which vary in size and quality. The structure has an exposed height of 4m, and a calculated total height of 9m, based on the assumption that the terrace reaches as far down as the foot of the ramp and staircase. The mud-brick terrace has only been exposed in a 2m wide test trench, but based on results of the geomagnetic survey, it can be deduced that the rectangular terrace was roughly 45 x 55m in 255 Suleiman & Taraqji 1995: 179, Fig. 32; Suleiman 2002a; 2002c; and: ARCANE database. 25° See: Quenet, this volume. Matthews 2003d: 109-115, Fig. 5.18-5.29. 258 Pfiilzner 2001: 146-153, 165 f., 169-176. 255 Dohmann-Pfalzner & Pfiilzner 1999: 39, Fig. 15; Pfalzner 2008b, 411 f.; Pfalzner & Dohmann-Pfalzner, in press. m Pfalzner 2008b, 412 f,; Pfiilzner & Dohmann-Pfalzner. in press. P. Pfalzner I ig. 53).261 The Mozan temple terrace is a "High terrace" in the S Mesopotamian sense and can be compared in aision to the known Early Dynastic examples at Ur, Uruk or Xippur.2"2 It provides clear proof that during the i y-millennium, the idea of a Ziggurat was not limited to Southern Mesopotamia but also existed in the Syrian I he high terrace was accessible via a ramp protruding over 20m to the S from the edge of the terrace. Its core is -•acted of mud-brick and it is trapezoid in shape, ranging from a width of 12m in the upper part to a maximum 1 ::. near the bottom. On the surface of the terrace were three low stone walls, forming a slight, step-like terrac-. .-. the upper side of the terrace block. This was intended to facilitate and direct the ascent from the ramp to the • r of the terrace. On top of the terrace was a temple building, Temple BA.26~' It is a single-shrine building with _ it-axis entrance and a stone altar block situated along the central-axis of the room. The room is particularly ■ _. with an interior of 9 x 16.5m. Buccellati reconstructs it with a gabled roof/63 however Pfalzner argues for aflat On the floor of the celia a rich pottery assemblage, typical of Period EjZ 3b was found,26 providing a precise : r terminus ad and ante quern) for the temple and the Stage II high terrace, on which the temple was built. r >m a general urban point of view and based on Early Dynastic comparisons in Southern Mesopotamia beid, al-Hiba, Khafajah) it is plausible that the huge oval temenos wall surrounding the high terrace already ted during Period EJZ 3 which correlates to the final ED II and ED III period of the S. However, the lowest : ivated level of the oval wall is associated with an EjZ 4 floor (Level C8), thus an earlier construction elate can-: vet) be proven archaeological!}'. The same is true for the stone staircase covering the ramp of the high terrace, Fig. 54: Beydar, Area F, Temple A, Phases 1-2, EJZ 3b (Dezzi Bardeschi & Stenuit 2007: PI. 2). -• Dohmann-Pfálzner & Pfálzner 1999: 29-39; 2002a, 172-179: Pfalzner 2008b: 410, Fig. 11. :': Pfalzner 2008b: 419-422. 263 Pfalzner 2008b: 425-428. 264 Buccellati 1988: 59-61, Fig. 9-11, ill. 6-13; Buccellati 1998: 18-21; Buccellati & Kelly-Buccellati 1995a: 389, Fig. 3; Kelly-Buccellati 1990: 125-129. •-' Buccellati & Kelly-Buccellati 1995a: 390, Fig. 3: 2005: Fig. 3; Buccellati 2005: Fig. 1. 266 Pfalzner 2008b: Fig. lb. '■" Buccellati & Kelly-Buccellati 1988: 65-66, ill. 12-13; 1995a, 390 f. 91 Architecture Fig. 55: Beydar, Area F3, Temples B and C, EJZ 3b (Suleiman 2007: PI. II). which plausibly already existed in Period EJZ 3. However, its earliest stratigraphic attestation is likewise in t nection with the aforementioned EJZ 4 floor (Level C8) (see below). A number of temples dating to Period EJZ 3b (Temples A, B, C, D, and E) were excavated at Beydar. The differ in dimensions but all belong to the same characteristic Beydar type.268 Temple A, 25 x 30m in size, accessed from an alley through a vestibule (32383) and a narrow corridor, which leads to a central square con yard (Fig. 54).265 The latter is paved with burnt bricks arranged in a herringbone pattern (32892). The courtyarc gives access to a large bent-axis cella (6682) 9 x 7.5m in size. The whole inner face of the cella's E wall is decorate. with a series of mud-brick niches, with every niche recessed twice. A small bench lines the base of the niched wall One of the niches is larger and deeper than the others and in front of it is a large, low podium. It is argued that this niche probably housed a cult image.2"" In the SE corner of the cella is a small adytum (32490), also with a niche-decotated outer wall. Another door opens from the cella into a small room (6899) with a toilet. There arc side rooms to the E, S and W of the courtyard and a group of rooms to the W of the cella, also accessible from the courtyard. These rooms, with the exception of a second toilet (12550), were used for economic purposes, mainh storage. This underlines the economic role of the temple, which seems to be closely related to the palace. This is also highlighted by the close proximity of the two buildings. South of Temple A is a group of two temples, Temple B and C, both strikingly similar in layout and fittings (Fig. 55). Temple B is entered from an alley up some steps, which lead into a small courtyard containing a flat basin and paved with burnt bricks laid in a herringbone pattern.2 1 The courtyard provides access to a staircase, a room with a toilet and a cella with a bent-axis entrance. The lattet has the same furnishing as Temple A: a double-recessed niched inner facade along one width, lined by a bench and bordered by a much deeper and broader niche, in front of which a large flat podium is situated, resembling a "cult niche". From the cella a second toilet room can be reached, and to one side is a room equipped with two short benches, probably used for storage. Temple C, situated parallel to Temple B, has a different kind of entrance. The courtyard is entered from, the E side through a vestibule, from where access is gained to a staircase.2 2 The courtyard is again paved with burn' bricks laid in a herringbone pattern. Tire temple is identical in la\ out and in approximate dimensions to Temple B: 268 Lebeau 2004. 2li> Bretschneider 2003: 93-106. Fig. 41-56, Plan 6-9; Bretschneider, Cunningham & lans 2(XP: 41-43, Fig. 1-2; Dezzi Bardeschi & Stenuit 2007; 53-61,PI. 1-2. Fig. 1-". m Bretschneider 2003: 98. m Suleiman 2003b: 169-170, Fig. 1-4, Plan 18-19; 200": 85-88, Fig. 1-5. Pi. HI. 2-2 Suleiman 2003b: 170 t. Fig. 5-8, Plan 18-19; 200": 85-88. Fig. 6-10. PI. HI. 181 P. Pfilzner from the courtyard there is access to a toilet room and to the bent-axis cella, with the same kind of double-recessed niched inner facade, a bench and a podium. Behind the cella one finds the second toilet room and a storage room. Temple C is additionally decorated by a niched outer temple facade along the adjoining alley. The high degree of standardisation in this EJZ 3b Beydar temple type is astonishing. Temple D, situated to the E of Temples B and C, again follows the layout of the Beydar temple-type. The courtyard is reached through a narrow vestibule and is paved with burnt bricks in a herringbone arrangement Fig. 56),2 3 From the small courtyard one can reach a toilet and two other rooms, whilst to the S it opens onto the cella. In the cella a bench and a podium are preserved close ro where the niched facade must have originally been, although this has now been completely destroyed. Two smaller rooms are again accessible from the cella. A newly discovered monumental temple at Beydar, Temple E, lies to the S of Temples B and C, and also dates to Period EJZ 3b (Fig. 57). It adopts features of the afore-mentioned smaller Beydar temples, but also reveals striking similarities with the large temple complexes FS and SS at Brak (see below).274 Temple E was probably accessible from the S through a courtvard paved with baked bricks again in a herringbone pattern, typical of both Beydar and Brak. From here, a huge hail (13471), approximately 20 x 16m in dimension, was accessed. On its inner E wall is a decorated niched facade, with a bench at its foot and a podium at its N end. These interior features closely resemble Temples A to D and, thus, adhere to the Beydar stvle of temple furnishing. The size of this room however is outstanding in comparison to the other Beydar temples but finds a close parallel in the central rooms of Temples FS and SS at Brak. At Brak they are interpreted as courtyards due to their size, however it seems, as Lebeau argues,275 plausible to reconstruct them as closed rooms, both at Beydar and at Brak. The vulnerable lime floors and the plastered facade make a strong case for the existence a covered space. Whilst it would have been challenging to roof the 16m space with wood, this should have been technically possible. The enormous 2m thick walls support the technical feasibility of roofing. From this large central hall a small room with a toilet (13455) in the NE corner of the building is accessible, again a typical feature of the Beydar temples. In contrast to the other Beydar temples, however, there is a succession ■jjP' j 'LOW WALL 4& INSTALLATION RECONSTRUCTION Fig. 56: Beydar. Area X. Temple D, EJZ 3b (Debruyne & Jans 200": Fig. 3). r' Debruyne & Jans 2007: 76-78, Fig. 2-3, 5-8. -"- Suleiman 2008: 13, Fig, 1, plan p. 15; 2 »09: I 3-15, Fig. 1-2, Plan p. 18: Suleiman 2010 n.d.; Lebeau 2010 n.d. r5 Lebeau 2010 n.d. Architecture TELL BEYDAR 2001-2009 AREA M EARLY JEZIRAH lllb SCHEMATIC PLAN Í.r_j WAU, UTES PH*£E «'-:...>-- i" ,v ;-v, , 3 RSCCKřTRyClJM ■_. RECJkSTKäTK^ L.J3 ftoop =*tfPWir^ F^DOH 3R03 3 THWWWlB Fig. 57: Beydar, Area M, Temple E, EjZ b (Suleiman 2010 n.d:, PI. I) (with kind permission of M. Lebeau). of two rooms to the E of the large hall, the first one (13496) accessible from the hall in a bent-axis arrangement whilst the second (13461) can only be reached from the first one. This parallels perfectly the characteristic roi arrangements in Temples SS and SS at Brak. Whilst at Brak, these two rooms are labelled ante-cella and ce/l.i. i: seems, at least in the case of Temple E at Beydar, to be more apt to designate them as adyta. On the opposite side of the central hall, in the W wing of the building, are two mote side rooms. This give the whole building a tripartite layout with the large hall at the centre and two side wings. The tripartite layout als structured the arrangement of the smaller temples B and C at Beydar, with the cella in the centre, but it can als be reconstructed in the (partly preserved) temples SS and FS at Brak (see below). As we saw already in the case the Tupkish palace at Mozan, the tripartite layout was also applied in palace architecture. This seems to have bee a universal concept in Early JZ architecture. The large temple complexes SS and FS at Brak, dated by the excavators to Btak Phase M,2 6 corresponding to th: Akkadian period (EJZ 4), might well be older than assumed, at least with regard to their date of construction. Ihe excavators alreadv assumed that the construction phase ofTemple SS, Level 5, dates to the transition from Brak Phas. L to M (= EJZ 3b to EJZ 4).2 It should be noted that Period EJZ 3b correlates with the ED IIIW early Akkadian (pre-Naram-Sin) phase.2"8 An EJZ 3b date for both temple complexes is particularly plausible given the striking similarities to the plan ofTemple E at Beydar,2 9 and furthermore would be in keeping with our knowledge regarding the historical importance of Brak, ancient Nagar, during the pre-Akkadian period, the time of the "Kingdom of Nagar" (Brak Phase L, ca. 2450 - 2350/2300 BC).280 Nevertheless, the established dating conventions will be followed and the two buildings will be presented in this chapter under the heading of EJZ 4 temple architecture (see below). The EJZ 3 temples at Khuera follow a completely different architectural tradition to those at Beydar. Here, the temple in antis type is most prominent in Period EJZ 3, of which no 3rd millennium examples have been recorded for the Khabur region of the Svrian JZ. Thus, there appear to be different sub-regions of architectural traditiom within the JZ. The most famous temple in antis at Khuera is the Steinbau I (Fig. 58). The oldest phase of the temple 276 Dates & Gates 2001a: 41. 73. 2 Ibid. 73.Thc arguments brought up by the excavators against an EJZ 3b date of complex SS (Oates & Dates 2001b: 39 It.;, mainly stressing the non-local chatacter ot the SS architecture, are not valid any more, especially with regard to new discoveries at Beydar Lsee above). 2"s Pfalzner 1998: 70, Fig. 1; Dohmann-Ptalzner & Pfalzner 2000: 191. rig. 2. 2 9 See: Lebeau 2010 n.d. 260 Oates & Oatcs 2001b: 380-383; Eidem, Finkel & Bonechi 2001: 100 f. 183 P. Pfiilzner Fig. 58: Khuera, Steinbau I, EJZ 3a & 3b (Orthmann 1990b: Fig. 9). precinct, Level 7, dates back to Period EJZ 3a (Khuera Phase IC). Levels 6-1 of the Steinbau I temple precinct date to the succeeding Period EJZ 3b (Khuera Phase ID)281. Ihe stone temple platform itself was built during Phase 7b, i.e. within Period EJZ 3a.2S2 Steinbau I temple was constructed on a high platform ("Kernbau") using large, sometime gigantic stone boulders.283 The structure is 26 x 14m in size with a height of 7m. Lower terraces built of stone and mudbrick are attached to the platform on its N, W and S sides. These were probably added latet to stabilise the construction. The temple proper on top of the platform, built using mud-bricks on a stone foundation, was not preserved, but could be reconsttucted on by studying the remains of foundations found within the podium. It was a classic temple in antis-type with a cella (interior of 15 x 10m) and zpronaos (5 x 10m) between two antis both 5m in length.284 The platform was accessible via a stepped ramp, obliquely attached to the platform on its E side. The tamp was initially made of stone (Stage 2) but later covered with mud-bricksteps (Stage lb).2S5 It is approximately 10m wide, thus hiding a large part of the platform's E front, and only 10m long. Adjacent to the temple platform's E and particularly N sides were dense clusters of rooms (Levels 7-l).2Sl5 The rooms were used for economic purposes, including storage. On the E fringe of the precinct, N of Steinbau II, was a bakery287. This clearly demonstrates the economic role of the temple, also evident at Beydar, where the four temples are surrounded by storage rooms and detached storage buildings. Steinbau II, located in Area B 30m to the SE of Steinbau I, also dates to Periods EJZ 3a and EJZ 3b (Khuera Phases IC and ID) (Fig. 59).288 This sttucture was formerly regarded as a temple in its own right.289 However, a re-study of the area showed that it actually functioned as a propylon-type entry building to the vast temple complex extending from there up to Steinbau I.290 Steinbau III, corresponding in date to Steinbau I and II (EJZ 3a and 3b), consists of a large stone platform with dimensions of 16 x 14m and 4m in height (Fig. 60). The platform was accessible via a monumental, 13m wide and 21m long staircase, of which 14 regularly laid stone steps have survived.291 The building has proven difficult to interpret functionally. Its initial excavators suggested that it was a temple on a high platform or perhaps 281 Meyer 2010c: 14. 232 Orthmann 1995: 34,43-46. Moortgat 1960a: 22-32; 1960b: 12-35; 1962: 22-42; 1965: 6-9; Orthmann 1995: 17-28. 281 Orthmann 1990b: 15-1", Fig. 9 Orthmann 1995: 24-28, Fig. S-7. 256 Moortgat 1960b: 20-31; 1962: 23-35; Orthmann 1995: 28-3". * Klein & Orthmann 1995: 73-75, PI lla-b; 12a. m Meyer 2010c: 14. m Moortgat 1960s: 25; Orthmann 1990b: 18, Fig. 10: 1990c: 253. 2911 Klein fie Orthmann 1995: 76; Fig. 32. Moortgat 1965: 9-11; 1967:4-8; Moortgat-Correns 1975:36 fF.; Orthmann, Klein & Liith 1986: 34-37; Orthmann 1990b: 19-20; 1990c: 251-253. P. Pfalzner numental grave.2-2 Orthmann originally argued for a reconstruction as a temple in antis with an open pro-3 Based on the re-interpretation of Steinbau II and on new comparative data, it now seems more plausible : 20 m) again lies at the heart of the complex. There is onlv one "/zwz»"-tvpe room (6), which opens onto the W side of rhe courtyatd. On the S side is a large gateway (1/2/13) with two door openings (2/ 13). It has one internal room (1) equipped with benches and bins, which was probably used as a waiting room. On the N side of the large courtyard is the temple, the courtyard facade of which has deep niches, but no platform or other typical furnishings similar ro complex SS. The excavators interpreted the main temple Room 43 as a courtyard. It has a size of 10.7 x >11 m. It is probable that the room was roofed, like the main hall of Temple SS. This is again supported by the thickness of the outer walls, especially the N wall (mote than 2m thick), and by the smoothness of the mud plaster floor/12 which would be unsuitable lor an open-air environment. Thus, the hall can be interpreted as the cella of the temple. From its 511 Oates &: Oates 2001a: 41-53, Fig. 42-59. 312 Matthews et. al 2001. 356 f„ Fig. 359 (based on the microstratjeraphic observations the floor proper docs not show rain disturbances, which are only visible in a layer on top of an ash deposit and date after the abandonment of the building). P. Pfalzner Fig. 69: Brak, Area FS, religious complex FS, Level 5, EJZ 4 (from Oates & Oates 2001a: Fig. 42). NW corner two small rooms, interpreted by the excavators as an antecella and cella, arc accessible. Again, as in the case ol Temple SS, the small size of the rooms and the installations do not support this interpretation. There is a tiny, flat podium and a circular bin in the first room (42). With the exception of a niche found high up on the side wall, there are no installations in the second room (41). The rooms should instead be interpreted as an adyton or a special treasury. The W wing of the FS complex contains storage rooms and a kind of reception room (20), very similar to those in the SS complex. Again, one can argue that additional economic and political tasks were undertaken in the FS complex, and that these were integrated with, or complementary to the assumed religious role of the building. However, it should be noted that even the possible religious role of the "temple" unit (41/ 42/ 43) is not substantiated by specific installations or finds. The evidence of EJZ 4 religious architecture at Khuera is scarce. The only clear example is "Steinbau" VI in the city centre. Its last phase of use, Phase 2a, can be dated to Period EJZ 4 (Khuera period IE). There are no major architectural changes between the EJZ 3b phases (see above) and Phase EJZ 4 of the temple, suggesting that there was continuity in the use of religious buildings between the two periods.313 At Beydar there is also one example of a Period EJZ 4 temple. Temple A was constructed immediately to the S of the Acropolis palace during Period EJZ 3b (see above). In the Early Akkadian period (EJZ 4a) the temple plan remained principally unchanged, with only minor modifications to the rooms around the courtyard (Fig. 70).31"' The cella retained its characteristic interior niched facade and presumably its cultic functions. During this phase a rich burial with two shafts was made below the floor of the cella.313 This suggests that the temple was probably used as an ancestor sanctuary, at least from Period EJZ 4 onwards. In the Middle Akkadian period (EJZ 4b) the cella was abandoned and replaced with a courtyard.316 Rooms of unclear function and no obvious relation to the cultic use of the building were constructed in the W part. In the Late Akkadian period (EJZ 4c), a new temple building was constructed. It consists of a nearly square room of 8 x 8m with a central platform, a podium, a basin on the W wall and a hearth in the NE corner.31 These installations are not like the cultic installations found in earlier temples, thus it is unclear whether this building maintained a religious function. "3 Orthmann 2002 5-7 m Bretschneider. Jans & Sukiman 2003: l-»9-15-». plan 12-13. Brets..... : " 99 it.. Fig. 1-j Architecture brick wall brick wall i installation jVHH| installation rzj iahuer phase Fig. 70: Beydar, Area F, square "temple" (Room 6334) of the Late Akkadian period, EJZ 4c (Bretschneider, Jans & Suleiman 2003: Plan 17). 5.5.5 Period EJZ 5 The only Period EJZ 5 religious building is the temple oval at Mozan/ Urkesh. Stage V of the temple oval is dated to Petiod EJZ 5. It is associated with three successive floors on the Max-Mallowan-Place abutting the temple staircase. The temple oval, consisting ol an oval perimerer wall, stone staircase and mud-brick terrace with a high temple, remained unchanged during Period EJZ 5, providing evidence for the uninterrupted religious nature of this important sanctuary ; the final part of the 3ra millennium. It provides a clear indication of continuity in religious structures and systems from Period EJZ 4 through to EJZ 5. 5.6. Storage Buildings 56.1 Period EJZ 1 The grill-plan structures recorded at Raqa" i, Levels 5-7 (see above) can be interpreted as substructures of storage facilities. They served to raise easily perishable goods from the ground and thus kept them dry.518 The grill-plan architecture does not tepresent public or communal building activities. Rather, it provides evidence of a small scale, household-based system of storage. It seems to have been characteristic i ic of EIZ 1 dome- stic structures. Another grill-structure, exposed on a much larger scale, was excavated at Ziyadeh on the Middle Khabur (Fig. 71).JW It dates to sometime between Periods EJZ ]-!,'''' although given the close comparability with at See the discussion above in section 5. 2 EJZ 1. Buccellati, Buia& Reimer 1991; Hole 1999; 2000. SM Hole 1999, Tb. 1: 2000. 191 P. Pfiikne; the structures ar Raqa'i321 an EJZ 1 date is most likely. It is the best-preserved example of this type of architecture, with walls preserved to a height of 2m. The building consists of three units: a grill-structure to the W, a "central building" with two rectangular rooms, and a square-shaped "eastern building".322 A mud-brick terrace, interpreted as a "loading platform" for the storage rooms, protects the building on the N side. The grill-unit contains five walls, each 50cm wide with 60cm spaces in between. An inner corridor connects the four narrow spaces via well-preserved arched doorways. The excavator reasons that the entire building served as a storage structure, due to the narrow spaces within the grill structures, making it unsuitable for habitation; the absence of doors in the well-preserved 2m high walls of adjoining rectangular rooms; and due to the complete lack of household installations within the rooms.323 The diversity of room types within the storage complex may reflect a diverse range of storage-related activities or types of stored goods. The existence of a tannur immediately to the W of the grill-structure324 hints at a possible connection of this large, multi-purpose storage facility with the adjoining house. 5.6.2 Period EJZ 2 The most prominent stotage structure of Period EJZ 2 is the Round Building at Raqa'i. It has been recorded in two levels, Level 4 dating to Period EJZ 2 (Fig. 72), and Level 3 dating to the end of Period EJZ 2 (Fig. 14a).325 It is completely preserved in Level 4 and sits at the centre of the small settlement. Houses surround the building, although these are mainly attested in Level 3 (see above). The structure is oval in shape with a diameter of 23m and is enclosed by a thick outer wall with onlv one entry, giving it a strongly fortified chatacter. The interior space is densely subdivided into 29 small chambers (Level 4), of varying size and layout, most of which served as storage rooms.326 At Atij, there are silos in both the N and the S areas of the settlement. The northern silos (Fig. 73) consist of three building units each housing small chambets of varying size, some of which show remains of corbelled roofs.32 Two of the small silo rooms (503, 504) have a T-shaped form due to the fact that the side walls arc only corbelled over two thitds of the room's length, whilst one third of the room has straight walls in order to allow access from the roof perhaps using a ladder.328 The three building units have, in contrast to the Round Building of Raqa'i, an overall rectangular outline. On the E side they are fortified by a thick wall, and a wide mud-brick platform, which is attached to the inside of the oval fortification wall surrounding the entire settlement.329 321 See Section 5.2 EJZ 1 322 Hole 1999: 269-272, Fig. 3-5. 323 Hole 1999: 269. 324 Hole 1999: 170, Fig. 6. For rhe dating compare Qucnet, this volume. 326 Curvers & Schwartz 1990: Schwartz & Curvers 1992: 406-410. •*r Fortin 1988b: 155-162. Fig. 18; 1990a: 222-232, Fig. 2-10. 328 See: Fortin 1988b: Fig. 19-20. 325 Compare Fortin 1995: Fig. 1 3.