S. Architecture Peter Pfälzner 5.1. Introduction Aside from pottery, architecture is the most ubiquitous form of EJZ material culture. As long as careful methodological principles of excavation have been applied, architecture can provide clear stratigraphic data. Thus, together with pottery, architecture forms the backbone of the EJZ relative chronology. Due to the large number of architectural remains in the JZ dating to the 3rd millennium, not all such evidence can be presented here. Instead, this discussion will focus on instances where the complete plan of a construction is known, or where diagnostic architectural features of an extended building have been exposed. These are the necessary prerequisites for studying the typology and function of buildings. Thus, the selection of ARC AXE architectural units is directed by the quality of exposed architectural remains with regard to their potential for typological and functional interpretation. Based on the methodological principles of the ARCANE project, the chronological ordering of the ar:r.;-tectural features and buildings depends exclusively on the nature of their stratigraphy rather than typolosici-inferences. The stratigraphic attribution is principally based on the information given by the excavators in the:: publications, including both preliminary reports and final publications, as well as the stratigraphic description; provided in the ARCANE Database.1 In addition, I refer in large parts of my chronological ordering to the art:;.; by Ph. Quenet (Chapter 2) on the stratigraphy and dating of sites. In cases where discrepancies arose between : he excavators' information in the ARCANE database and the article by Ph. Quenet, the latter was followed in —. ■■:.-: cases since his chronological conclusions are based on a thorough and in depth comparison of pottery and other inventories from all relevant sites. Quenet's contribution on stratigraphy in this volume is an invaluable roc; ::: understanding EJZ chronology. My views on the dating of levels and on the typology of the architecture are also based on my own research the JZ, primarily at the sites of Bderi, Khuera and Mozan. This naturally leads to a strong focus on the archht;-tural features of these three sites in this article, based on my detailed understanding and own evaluation or :h.-.; architectural contexts.2 This also explains the strong emphasis on private houses and domestic insta!li:::.v -based on my book "Houses and Household"3 - as well as public spaces and streets in this contfibution, e-ri:.i. as explored at Khuera and Mozan. In this article, the architecture has been divided into five functional categories: 1. Fortifications; 2. H.. : and domestic installations; 3. Palaces; 4. Temples and ceremonial buildings; 5. Storage buildings. The r..-.:=r._ presented accordingly and within each category the buildings are discussed in chronological order, sub-;. :.; into the EJZ phases. Despite the large amount of excavated architectural remains in the Syrian JZ, there are so far n: ;. - : plans or larger exposures of any buildings dating to Period EJZ 0. This period is (with the exception c:::: : . ... category) not therefore treated as an independent period in each functional category. The chronologlr i. -:.-in each functional category will thus be: EJZ 1 - EJZ 2 (EJZ 2 and final EJZ 2 will be handled together - 1 "Z (subdivided into EJZ 3a and 3b only for houses) - EJZ 4 (sub-periods EJZ 4a, 4b, and 4c are treatei ..-to the scarcity of architectural data) - EJZ 5. It has already been noted that the quantity of excavated architectural structures for the EJZ re.-..; . :_z- so that only a selection of the material can be presented. On the other hand, as will be seen, there ire . - . r._ir, insecurities and lacunae in the data for this period. For Periods EJZ 3a and 3b, a lot of material Ls :_-_c: all functional building categories, whereas evidence for various categories is more restricted in zs-'.:re»s: Volume.» 2 ar.i ? irr ilreii - kTrarti .r . -Deckers et al. 2010), whilst Volume 1 (architecture mi »:ri:icrirhv -x'Z rr.r.rec er- •: 3 Pfalzner 2001. P. Pfdlzner 5.2. Fortifications 5.2.1 Period EJZ1 The development of fortifications and fortified settlements in the JZ Region began during Period EJZ 1, Evidence recently retrieved at Khuera demonstrates that the inner city of the "Kranzhiigel" was founded during Period EJZ 1 (Period Khuera IA). It has been assumed that this was already fortified.4 The exposed part of the inner city wall is built of mud-bricks and has a width of 4.2m. It is connected to a possible "tower" or "bastion", but it is-not yet clear from the stratigraphy whether the city wall already existed during this early period.5 Clearer archaeological evidence of early fortification walls is available from smaller sites in the Khabur region, especially in the Middle Khabur valley. At Atij there is an oval perimeter wall on the summit of a very small tell6 (Fig. 1). This "mur d enceinte" is attributed by the excavators to Period EJZ 2,7 but is dated by Quenet to Period EJZ l.8 It is built on virgin soil,9 thus demonstrating that the settlement was fortified from the beginning of its existence. The 2.5m wide wall is entirely constructed of mud-brick up to a preserved height of 4-5m10. Although the wall was only excavated in three places - in the -, the middle and the S of the tell - it was possible to infer that it surrounded the whole of the former settlement. It once encompassed a very modest total settlement area of only 80m (N-S) by 50m (E-W), and was regularly curved to form an overall elongated oval shape.11 It must have been primarily intended for the protection of stored goods rather than of the people living in this tiny settlement. The contemporary fortification wall at Kashkashok III is similar in shape and was exposed in a wide stretch of Area A (Fig. 2). It is referred to as the "Rounded Building". It was constructed in Level AIX (EJZ 1) and seems to have continued in use daring Period EJZ 212. The 1.2m to 2m wide wall is built of mud-brick13. Like the fortification wall at Atij it is smoothly rounded in shape, but in contrast to the latter it is not purely a single-line wall. It consists of a single wall line in the E part, but spreads into two parallel walls in the N part of the site, which enclose seven small, casemate-like rooms. The wall is not excavated in the W and S part of the tell, but seems to have been combined - at least in later phases - with secondarily constructed, elongated room units in order to form a completely closed, fortified circle or oval. It must have protected open courtyards and houses located inside. The total extent of this fortification - with a reconstructed diameter of roughly 65m - is very similar to the one at Atij and seems to have surrounded the whole site. Another EJZ 1 fortification wall was detected at Kneidij, in the oldest Level XVI at the site.14 This mud-brick wall (Wall 559) was only detected in a small sounding on the SE side of the settlement15. It was constructed on virgin soil, demonstrating that the settlement at Kneidij was also fortified from the beginning of its existence. In the later Period EJZ 2 the wall is attested to have surrounded the entire settlement (see below). It is not yet possible to demonstrate whether this was the case for Period EJZ 1 due to the limited exposure. In summary, the evidence clearly proves that many EJZ settlements, from larger urban (e.g. Khuera) to small rural sites (e.g. Atij), were fortified from the very beginning of their existence in Period EJZ 1. 5.2.2 Period EJZ 2 At Khuera one can observe the successive development of the fortification system. Whilst the inner wall possibly had already been erected in Period EJZ 1 (see above), the outer city wall was added in Period EJZ 2 (locally labelled Khuera Phase? IB), as evidenced in Area Z.16 Through this additional construction, the typical two-crowned "Kranzugel" shape of the city was created; it thus appears to have been a two-stage process.17 The outer 4 Meyer, Chapter 4, this volume, and Meyer 2008c; 2010a:25. 5 Falb 2010:95-97, Fig. 8,36. 6 Fortin 1994c:377; Fig. 15; 1995:37-41; Fig. 12-14. 7 See ARCANE database. 8 Quenet, Chapter 2, this volume, (based on his analysis of the Atij pottery; pers. comm..). 9 Fortin 1994c: 377; 1995:40. 10 Fortin 1994c:377; 1995: 37, Fig. 12. ^ 11 Fortin 1995:Fig. 13. 12 See ARCANE database; and Quenet, Chapter 2, this volume. 13 Suleiman & Taraqji 1995:172-183; Fig. 32. 14 Dates according to Quenet, Chapter 2, this volume. 15 Klengel-Brandt, Kulemann-Ossen & Martin 2005: 14, Taf. 5 (section 12/2). 16 Meyer 2006b: 182; 2010a:25. 17 Meyer, this volume, Chapter 4; Meyer 2010c: 24 f.; 2010c: 181. 138 Architecture Fig. 1: Atij, fortification wall, EJZ 1 (Fortin Fig. 2: Kashkashok III "Rounded Buiidir 1995: Fig. 13). EJZ 1 (InventoryJZ012, Tb. 2 . city wall, the oldest phase of which, based on the stratigraphy, clearly dates to Period EJZ 2, is consrr mud-bricks without a stone foundation. It is 5.5m wide and is preserved to a height of 3m. The rortir. protected on its outer side by constructed compartments filled with pebbles, probably to protect the rr. city wall from water damage18. Period EJZ 2, thus provides a date for the establishment of a "Kranzhugel" structure at this site." with a width of 4.5m, was made of pise &nd might have been the foundation of a destroyed mud-brick >u: ture.20 The Beydar city wall originally had seven gates. These have not been excavated but have been : from the topography of the site and through references in the Beydar administrative texts. There is some that the NE inner city gate (see below, EJZ 3) was already in use during Period EJZ 2.21 At Kneidij, the fortification is not augmented as it is at Khuera, but continuity of the existing ror::r.; = tem from EJZ 1 to EJZ 2 is observable. For building Level XIII (EJZ 2) the city wall has been exposed o~*-of 14m N of Building Complex N22 (Fig. 3). It was built of mud-brick and had a width of 2.4m. On :hc : is protected by a glacis of equal thickness as the wall and reinforced by obliquely laid mud-bricks. This wil surrounded the whole 3ha settlement, which in this instance, was characterised by a densely populates ij multi-room houses and storage facilities (see below). In contrast to earlier fortifications at Atij or Kashxif and the round city wall of Khuera, the exposed part of the wall at Kneidij forms a straight line. This. to the view of the general topographical layout of the settlement, suggests that the fortification svsterr. v rectangular or at least polygonal in form. Continuity of fortifications is also attested at Kashkashok III. The documented evidence so fa: :r.z;c the EJZ 1 "Rounded Building" was used continuously during Period EJZ 2, with buildings being added : on in Levels A VIII to A VI (see above and Plan Fig. 2).23 There is a marked increase in fortified settlements during Period EJZ 2. An example of a r.tr--.-site is Bderi. Here, a city wall and a city-gate existed from the very beginning of the EJZ 2 sequence i: 18 Meyer 2010e: 173-176. 15 Bluard 1997: 179-192; Bretschneider 1997: 193-208. 20 Bluard 1997: 181: Bretschneider 1997: 194, Fig. 3. 21 Mikno & Rova 2008: 588, Fig. 2-3. 22 Klengel-Brandt, Kulemann-Ossen & Martin 2005: 33 f.. Taf. 26, 40. 23 See: ARCANE database: and Quenet, this volume, Chapter 2. 139 P. Pfäizner Fig. 3: Kneidij, fortification wall, Level XIII, EJZ 2 (Klengel-Brandt, Kulemann-Ossen & Martin 2005: PI. 40). in Level 27 (Fig. 4).24 It directly overlies a ditch belonging to Level 28 (EJZ 1). This most likely served to drain off water from the regular inundations of the Khabur, rather than for fortification purposes. This new city wall existed throughout Period EJZ 2 (Levels 27-21) and continued to be used during the succeeding EJZ 3a period. The fortifications were therefore important in structuring the urban development of this small 6.8ha town. The wall, measuring 2m in width, is constructed of mud-bricks overlying a similarly built foundation of mud-bricks.25 A glacis is attached to the outside, entirely built of rammed earth. It is also 2m wide at the base and served to protect the city wall from being undermined by water erosion, especially during Khabur inundations. Furthet, the glacis existed from the beginning of the city wall's construction (Level 27) and forms an integral part of the wall. On the interior the city wall is strengthened by mud-brick buttresses placed at regular 2m intervals. The voids between the buttresses came later on in Period EJZ 2, enclosed by an attached wall creating tiny casematelike rooms between the buttresses. The city-gate at Bderi, best attested during Phase 25, is the oldest Early JZ period example so far excavated. The gateway is 3m wide and both sides are coated by a large limestone orthostat. Although their surfaces have not been smoothed and they are not exactly rectangular, the use of the stone orthostats must be seen as a precursor of the later Syrian orthostat-gates of the second millennium BC. Thus, the Bderi EJZ 2 city gate is a typologically important piece of evidence for the early development of gates in Syria. Special attention has been devoted to the circumvallation of Leilan, regarded by its excavator Harvey Weiss as important to the understanding of mid 3rd millennium state formation in the Syrian JZ2S. On the acropolis of Leilan a city wall has been identified on the basis of two large mud-brick walls (Walls A and B, positioned perpendicular to each other), which were mainly recorded in the sections of Trench 44W12/ X12 and could be traced for a length of 10m.27 Wall A was approximately 2.7m wide, but there is no clear indication about the extension and outline of this fortification system.28 It was erected at a specific point in time between Levels 16 and 15, thus dating to the end of Leilan Period Hid (= Levels 17-15), which corresponds 24 The Bderi city wall was dated in preliminary reports to Level 25 (Pfälzner 1986-87a; 1986-87b; 1988), but study of documentation during preparation for the final publication revealed that the wall existed from Level 27 onwards. 25 Pfalzner 1986-87a; 1989-90; 1990; 1992-93. 26 Weiss 1990a; 1990b; 1990c. 27 Calderone & Weiss 2003: Fig. 1, 2. 28 Ibid. Fig. 4, 5. 1-tO Architecture 9 1 : lj 5 ii 19 ip ! j 2^42 3 * 5 ( gff iTO-W1 rh; IMS w*1" ECH1CHT; 25 (STADTMAuea! WW SUCH Fig. 4: Bderi, city-wall and city gate, Level 25, EJZ 2 (Pfalzner, 1989-1990: Fig. 19). to the end of Period EJZ 229. The Leilan city wall remained in use during the subsequent Leilan Ha period (EJZ 3a and 3b).30 Excavations in 2002 at the N city gate belonging to Leilan's lower town revealed that both the gate and Leilan's lower city wall were erected in Period Leilan Hid (final EJZ 2).31 The mud brick city wall (Wall A) is 3m wide in this phase and is connected to an earthwork, which probably served as a kind of glacis32. This evidence demonstrates that the fortifications of Leilan's lower town also date to Period EJZ 2 and are thus contemporary to the acropolis (upper town) wall. The defensive wall at Rad Shaqrah was - according to Quenet33 - probably already built in EJZ 2, at least in its initial building stage. There were later extensions, probably during Period ED III/ EJZ 3.34 The wall was 3.9m thick and built of mud-bricks on a high foundation of basalt boulders (Fig. 5a).35 The exterior was protected by a glacis made of earth, broken mud-bricks and basalt boulders, with an outer face of tightly packed stones (Fig. 5b). This glacis was at least 5m high and 6m wide at the base36. This is one of the most impressive EJZ defensive systems discovered so far. It enclosed the tiny settlement of Rad Shaqrah, a small town with densely packed housing.37 It is roughly polygonal in shape, extending not more than 100m N-S and 50m E-W.38 A discussion of EJZ 2 fortification systems requires a mention of the fortified storage building of Levels 3 and 4 at Raqa'i. The site's fortified "Round Building" is, however, situated within the settlement with houses and other domestic areas grouped around it. Thus, it was not intended for the overall defence of the settlement. This important structure will therefore be discussed in connection with EJZ storage buildings (see below). 29 Ibid.: 198. 30 Ibid.: 198-201. - 31 Ristvet 2007:185-192, Fig. 6. 32 Ibid.: 189 f. Fig. 4. 33 Quenet, this volume, Chapter 2 (as opposed to a ED III/ EJZ 3 date given in the ARCANE database) 34 Bielinski 1993: 127, 1996: 162. 35 Bielinski 1992: 80, Fig. 1. 36 Bielinski 1992: 80; 1993: 119,123, 125 Fig. 1; 1994: 157-159, Fig. 2. 37 Bielinksi 1995: 117. 38 Bielinksi 1994: 155. Fig. 1. 141 P. Pßlzner Fig. 5a: Rad Shaqra, defensive wall, EJZ 2 - EJZ 3, plan of Area A with wall and glacis (Bielinski 1992: Fig. 1). Fig. 5b: Rad Shaqra, defensive wall, EJZ 2 - EJZ 3, section of Area B3 with three layers of stone glacis (Bielinski 1994: Fig. 2). At Rhazne a thick fortification wall was discovered surrounding a complex of buildings, which have tentatively been interpreted by the excavators as religious (due to the occurrence of niched facades) and administrative.39 The 4.3m thick wall (Wall 17) is described as being built of rammed earth plaques and is circular in layout (Fig. 6a).40 In the interior of this fortification, at a distance of 18m from the outer wall and not attached to it, an originally freestanding mud-brick tower (Room 37) was exposed, which was preserved to a height of 8m.''1 The exact chronological relationship and developmental sequence of these multi-layered structures cannot be ascertained, because the buildings were not distinguished according to single level plans. Although the stratigraphy and chronology of the site is far from clear, the main Levels III-I have been attributed to the EJZ 2 by Quenet.42 Thus, Khazne can be regarded as a strongly fortified settlement of Period EJZ 2 with probably specialised functions that remain to be assessed. 5.2.3 Period EJZ 3 A precise insight into how EJZ 3 city-walls were constructed can be gained at Khuera. The outer city wall was exposed in Area P, indicating that it existed in building Phase 2 (TCH Period IC = EJZ 3a) and building Phase 1 .TCH Period ID = EJZ 3b). The defensive system consisted of the wall proper on the interior, a glacis, a forewall 20 22 Fig. 6: Khazne I, fortification system around "monumental complex", EJZ 2 (Munchaev, Merpert - Amirov 2004: PI. 6). 53 Merpert & Munchaev 1999b: 119 f., 123; Munchaev, Merpert & Amirov 2004. 40 Munchaev, Merpert & Bader 1993: 163, Fig. 2. 41 Munchaev, Merpert & Bader 1993, 162 £, Fig. 2-5; Merpert & Munchaev 1999b: 120-122. '- Quenet, this volume, Chapter 2 142 Architecture ("Vbrmauer") and on rhe outer edge, a ditch Fig. ~V? The massive city wall, preserved to a height of nearly "m, measures 5.5m in width and was constructed of mud-brick"". The only 1.5m thick pre-waU was erected at a distance of 6m to the N of it, and was constructed (.during a eariier phase) of mud-bricks on top of a foundation of large limestone boulders. Direct.lv in front of the pre-wall was a kind of deepened ditch. The glacis, on the other hand, was erected at a later phase between the wall and the pre-wall. It was made f compacted earth and stabilised using a line of mud-bricks. This complex fortification system is a continuation of the outer city wall erected during Period EJZ 2 (see above). It remained in use throughout Periods EJZ 3a and 3b and is an excellent example of a typical city wall during the main phase of 33 Schwartz & Curvers 1992: Fig. 8; 1993-94: 250 f., Fig, 74. 54 Bielinski 1992:81; 1993: 121; 1995: 111 f. " Bielinski 1992: 81 f., Fig. 1; Pfalzner 2001: 313, Taf. 35 149 P. Pfalzner Fig. 16a: Rad Shaqra, Area A, house with interior Fig. 16b: Rad Shaqra, Area B, single-room houses with buttresses, EJZ 2 (Pfalzner 2001: PI. 35). interior buttresses, EJZ 2 (Bielinski 1995: Fig. 1). be reconstructed with flat roofs. Bielinski designated these types of houses "arcaded houses".96 These housebuilding techniques are characreristic of Period EJZ 2.97 The installations and finds in Locus A/1; a central hearth, a gypsum-plastered basin, and storage vessels between the buttresses,98 clearly hint that this room was a multi-functional living area ("nuclear room").'9 A second smaller room was added, so that this house formed a two-room house.100 Other examples of such houses with buttresses, excavated in Area B,101 were definitely single-room houses (Fig. 16b). Ample evidence of EJZ 2 houses was brought to light at Kneidij. The best examples pertain to Level XIII (EJZ 2).102 In contrast to the single and two room houses at Atij and Raqa'i, those at Kneidij are much larger multi-roomed complexes. Eleven house complexes could be distinguished (Complex D, E, F, G, H, J, K, L, M, N, O) in the middle and W part of the excavated settlement area. They are arranged in blocks either side of a straight street and beside an open area to the E (Fig. 17) wheteas the houses at contemporary Raqa'i are much more randomly scattered (Fig. 14a and 14b). Thus, both the house plans and the ovetall arrangement of the domestic areas at Kneidij difter considerably from the contemporary (and nearby) settlements of Atij and Raqa'i. This might be due to a fundamental difference in the socio-economic status of the settlements, the latter being tiny villages with communal storage,103 while Kneidij can be described as a small walled town. These two types of socio-economic settlement organisation, each with their distinctive concepts of house design, coexisted in Period EJZ 2. The houses of Kneidij are built in a very irregular layout without any generalised plan. Complex E, consisting of eight rooms, is a characteristic example and has been completely excavated104 (Fig. 18). The house has three inter-connecting, irregular courtyards (VII L, O, G) in the centre of the complex, containing a latge bread oven. % Bielinski 2005a. * Quenet, this volume,argues that the Rad Shaqrah fortification wall and attached structures date to the EJZ 2, rather than the ED II as originally assumed by the excavators. 98 Bielinski 1993: 121. 99 Pfalzner 2001: 314. 100 Bielinski 1993: 23. 101 Bielinski 1995: 111, Fig. 1. 102 Klengel-Brandt & Kulemann-Ossen & Martin 2005: 18-34, Taf. 26-41. 103 Pfalzner 2002b; 2008a: 168-169. m Ibid. 24-26, Taf. 34. 150 Architecture TALL KNHDIÖ/ NO-SYRIEN SCHEMATISCHE« ARCHITEICTURPLAN BAUSCHICHT XIII Fig. 17: Kneidij, Level XIII, house complexes (in the middle and W part of the settlement) EJZ 2 (Klengel-Brandt, Kulemann-Ossen & Martin 2005: PI. 26). Fig. 18: Kneidij, Level XIII, house of Komplex E, EJZ 2 (Klengel-Brandt, Kulemann-Ossen & Martin 2005: PI. 34). Room VII, containing the typical installations of a central hearth and wall benches, can be identified as the main living room ("nuclear room") of the house. It is directly accessible from Courtyard VII L. Adjoining the nuclear room is Room VII M/K, with inner buttresses, possibly for arches or quarter-arches. This arrangement is strongly reminiscent of the EJZ 2 architecture at Atij (see above). Attached to the courtyards is a semi-subterranean room with grill-plan structure (Room VIII/J), making it suitable for storage. This is a clear continuation of the typical grill-plan architecture of the earlier Period EJZ 1, which confirms that its use persisted into Period EJZ 2 as already seen at Raqa'i. In conclusion, we can note a combination of different architectural traditions 151 P. Pülzner % % t 2 10 Fig. 19: Khuera, Area K, Level 9b (period TCH IB), Houses, EJZ 2 (Pfälzner & Dohmann-Pfälzner 2002: Fig. 3). in the houses of Kneidij, their integration into larger units reflecting complex domestic activities and household structures. The houses in a large urban centre such as Khuera represent yet another type of socio-economic arrangement. The earliest completely excavated houses at Khuera are attested in Period TCH IB, which corresponds to EJZ 2. In Area K of the upper city, close to the central plaza these EJZ 2 houses have been exposed in Levels 11 to 9105 (Fig, 19). The houses are aligned, as at Kneidij, along a straight, narrow alley (Gasse A). From here, the houses are accessible either directly (Houses la, Ha, lib) or through a vestibule (Houses lb, XVIII, XIX) leading into an elongated or square courtyard. The bread ovens ifananir) are located in small shelters attached to the courtyard. The main room of the houses is located in the rear part of the house complex, directly accessible from the courtyard. It is usually equipped with a rounded central hearth, a mud-brick podium (to be used as a house altar106), and benches along the walls, thus forming a typical "nuclear room". One to three smaller, additional rooms, mostly located along the alley in the front part of the house, function as areas for storage or food processing. In summary, the EJZ 2 houses of Khuera are essentially similar to those at Kneidij, although tend to be smaller in terms of the number of rooms. This can probably be understood as a consequence of the dense urban living conditions at Khuera. 5.3.4 Period EJZ 3a In Period EJZ 3a a completely new concept of house design appears. This type has been designated "allotment houses" ("Parzellenhduse)-")107, defined as houses built on a regular, rectangular plot of land with fixed dimensions. Plots are usually 6m, 7.5m, 9m, 12 m, or 15m in width. The smaller dimensions of 6m and 7.5m are the most frequent. This standardisation of house plots is explained by a process of planned urbanisation carried out under the supervision of central institutions. These institutions must have carried out the measurement, division and allocation of house plots. This strategy was intended to ensure a planned expansion of domestic living quarters as well as attract people to urban centres.108 Thus, "allotment houses" are characteristic of this period, when a policy of urban expansion was at its climax in the Syrian JZ, i.e. in Period EJZ 3a10'. The standardisation of house plots was linked to standardisation in house plans. The exact reasons for this are difficult to determine: either the standard house plans resulted quasi automatically from the allotment of regular house plots, especially where common building traditions existed, or this indicates that at least some of the houses where erected by the central urban planning institutions themselves. Dohmann-Pfälzner & Pfälzner 1996; Dohmann-Pfälzner & Pfälzner 2002b: 6-8, Fig. 3. Pfälzner 2001: 169-176. Pfälzner2001;378f. Pfälzner 2001: 378 f., 395-401; 2002 = : Dohmann-Pfälzner & Pfälzner 1996: 10-12; 2002b, 13 f. Pfälzner 1997: 249-254; Pfälzner 2010. 4-6. Tab. 2; in press; Pfälzner & Dohmann-Pfälzner, in press. 152 Architecture Fig. 20: Khuera, Area K, Levels 4-5, organisation of plots for "allotment houses' EJZ 3a (Pfalzner 2001, PI. 60). Khuera is one of the most prominent sites, where urban expansion based on the construction of "allotment houses" has been found. House remains of Period EJZ 3a have so far only been excavated on a broad scale in Area K ("Kleiner Antentempel"). Period EJZ 3a levels (Levels 7-9) have also been reached in Area H ("Hanserviertel") but no complete house plans have been revealed due to the limited size of the trench on the slope.110 Area K is therefore the main reference for EJZ 3a houses at Khuera. Levels 8 to 4, datable to Period TCH IC (= EJZ 3a) provide a continuous developmental sequence of "allotment houses". The oldest examples of "allotment houses" (Houses III and IV of Area K) were recorded in Level 7, which belongs to an early phase of EJZ 3a111. The extended domestic quarter, separated from the city centre by a terraced wall, is accessible through two straight parallel alleys.112 Overall, 13 (completely excavated) house plots can be distinguished within the residential quarter113 (Fig. 20). This includes the so-called "Kleiner Antentempel", which has been re-interprered as a domestic structure (Flouse Ha) during Period EJZ 3a (Levels 5 and 4).114 Most of the house plots in the E strip along Alley A have a standardised width of 6m (Houses K la, Ila+b, V, VII). The larger dimension of 7.5m is mainly attested in the W strip along Alley B (Houses K III, IV, VI, VIII, X, XI), where one house with a probable width of 9m is also found (House IX). In Area K at Khuera, the "allotment houses" differ in plan and internal layout. Typical plans characterise Houses K III and KIV (Fig. 21).115 In Level 5c these two houses are accessible from Alley A through an entrance corridor, equipped with a drain to evacuate water from the inside. A large courtyard beyond the corridor gives access to all rooms. The main room (Room 41, Room 126/127), which in these two cases is subdivided by internal walls, is situated immediately to the side of the enttance corridor, along the front side of the house and parallel to the alley. This is a very characteristic feature of EJZ 3a (and 3b) "allotment houses", observable in most houses of Areas K and H at Khuera and at other sites. This architectural principle did not exist in the previous EJZ 2 period, when the main rooms of the houses in Khuera (Area K) were situated at the rear of the house, far away from the alley (see above and Fig. 19). The installations within the houses are related to domestic activities and are thus usually found within these types of houses. There is nearly always a hearth in the centre of the main room, and often benches, gypsum basins or other installations, thus defining the room as a multifunctional core-room of the household, a "nuclear room".116 Grinding tables are located in separate small rooms, which are used as "grinding rooms".117 Bread ovens 110 Klein 1995: 106-108, Fig. 58-61. 111 Dohmann-Pfalzner & Pfalzner 1996: 8-11. 112 Dohmann-Pfalzner & Pfalzner 1996: 11 f. 1,3 Pfalzner 2001: 334-345, Taf. 60. 1M Pfalzner 2001: 337 £, Taf. 61 bottom. m Pfalzner 2001: 339 f, Taf. 62; Dohmann-Pfalzner & Pfalzner 1996: 8-9, Fig. 4. 116 Dohmann-Pfalzner & Pfalzner 1996:10; Pfalzner 2001: 149-153. 117 Pfalzner 2001: 139-146. 153 P. Pfalzner (tananir) are primarily located in or close to the courtyard and often within tiny separated chambers.118 These sets of installations are characteristic for most of the "allotment houses" at Khuera, Bderi, Melebiya and other sites, both in Periods EJZ 3a and EJZ 3b. At Bderi, "allotment houses" are attested in Levels 17 to 14, i.e. during Period EJZ 3a. They have a surprisingly similar layout to those at Khuera. The best example is a small house measuring 42m2, House XVII, which was found in Level 17 (Fig. 22)."' It is a 6m wide "allotment house", accessible from the alley through an entrance corridor with a water-drain. The "nuclear room" (Room FL) is located in the front of the house, beside the entrance corridor and parallel to the alley. It is equipped with a round central hearth. Its inventory hints at multiple domestic activities, such as food preparation, cooking, bread baking (during bad weather), storage of food and tools, heating, family gatherings, and also textile production. There are only two further rooms; these are small and were used for storage. Access from the courtyard to the "nuclear room" was only possible via these tiny rooms. The other fully excavated EJZ 3a "allotment houses" at Bderi, Houses XIV and XXI (both Level 14) have a different layout (Fig. 23). By contrast, the main room is not at the front of the house, however they do exhibit similar structural and functional characteristics.120 It is particularly interesting to not.e that House XIV had already been constructed during Level 20, which dates to the previous EJZ 2 period, and remained in use, with several internal modifications, until Level 14. Thus, it is the oldest know "allotment house". This means that the principle of "allotment houses" must have been created during Period EJZ 2, but its wider application throughout the Syrian JZ was not achieved until Period EJZ 3a. At Melebiya, Period EJZ 3a houses were only exposed in Area C, in the N part of the settlement. They date to Melebiya Phase 3, which corresponds to EJZ 3a.121 Only House C2 was completely excavated122. It was a very small house with only two (or probably three) rooms. It is 6 x 6m in size and, thus corresponds to the smallest 118 Pfalzner 2001: 146-149. "' Pfalzner 2001: 294 f„ Taf. 21, Tab. 66-67. 120 Pfalzner 2001: 293-295, Taf. 19, 20, 22; Tab. 64, 65, 68, 69. 121 See ARCANE database, and Quenet, this volume, In Lebeau 1993 this level was dated to the Early Dynastic II period; this was later corrected. 122 Lebeau 1993: 101, PI. 98-102; 1996a. 154