Anatolica XLI, 2015 EXCAVATIONS AT ZIYARET TEPE, DIYARBAKIR PROVINCE, TURKEY, 2011-2014 SEASONS Timothy Matney, Tina Greenield, Kemalettin Köroğlu, John MacGinnis, Lucas Proctor, Melissa Rosenzweig, and Dirk Wicke Abstract his article presents the results of excavations at Ziyaret Tepe, the Late Assyrian city of Tušhan in the Diyarbakır Province of southeastern Turkey during the summers of 2011-2013, as well as from a study season in 2014. Excavation in nine operations is briely summarized, and the preliminary results of zooarchaeological and archaeobotanical studies in three operations are presented. Major public buildings of the Late Assyrian period (c. 882611 BC) were recovered in Operation AN (palace), Operation W (administrative building), and Operation Y (city fortiication), while an exposure of domestic architecture was revealed in Operation K. Also of importance from these ield seasons was the documentation of a Late Roman, or Late Antique, occupation in Operations T and U. Zooarchaeological evidence from earlier excavations in Operation K reveal the subsistence practices of commoners during the Late Assyrian period. Likewise, the use of plants for human food and animal fodder are discussed for the Late Assyrian (Operation Q, the city gate excavated earlier) and the Late Antique (Operation T, domestic housing) periods. hese combined reports outline the importance of animal husbandry, as well as agricultural production of grain, as key economic aspects of the Late Assyrian settlement, and complement existing cuneiform documentation. Introduction his paper represents the inal installation in our preliminary report series in Anatolica covering the excavations at Ziyaret Tepe, the Late Assyrian city of Tušhan, in the Diyarbakır Province of southeastern Turkey. he 2011, 2012, and 2013 summer excavation seasons are discussed, as are some results from an additional study season undertaken in June-August 2014 to complete the recording of small inds, pottery, animal and human bones, archaeobotanical remains, and microdebris samples stored in our expedition depots.* For a more detailed de* Prof. Timothy Matney served as the director of the project. Our senior staf included: Dr. John MacGinnis of Cambridge University, Dr. Dirk Wicke of the University of Mainz, and Prof. Dr. Kemalettin Köroğlu of Marmara University. We would like to acknowledge the kind help of the directors of the regional Diyarbakır Archaeological Museum: Nevin Soyukaya (2011-2012), Mehmet Akif Bilici (2013), and Mehmet Eneze (2014). Likewise, we were assisted by a number of able and helpful government representatives: Ahmet Durman and Nuray Çırak (2011), Murat Aktay (2012), Esma Bedirhanoğlu (2013), and Leyla Ay (2014). Our thanks go out to all their eforts and hard work on our behalf. Funding for these seasons was generously provided by: the National Endowment for the Humanities (Grant No. RZ-50721-07), Deutsche Forschungsgemeinschaft (DFG), Johannes Gutenberg-Universität Mainz, the McDonald Institute of Cambridge University, the National Geographic Society (Grant No. 9290-13), the Tekfen Foundation, Timothy Matney et al. 2 scription of the site’s topography, its historical signiicance, and the results of previous seasons, the reader is directed to earlier reports in this journal (Matney 1998; Matney and Somers 1999; Matney and Bauer 2000; Matney et al. 2002; 2003; Matney and Rainville 2005; Matney et al. 2007; 2009; 2011). A number of important reinements and additions can now be made as result of our most recent excavations and analyses and these early reports should be understood as preliminary in nature. Fig. 1. Location of Ziyaret Tepe in the upper Tigris River valley of southeastern Turkey. he Site of Ziyaret Tepe Ziyaret Tepe sits within a fertile area of the upper Tigris River alluvial loodplain, approximately 20km west of its conluence with the Batman Su and just east of the modern city of Bismil (Fig. 1). he site falls within the impact area of the large Ilısu hydroelectric dam soon to be completed downstream and our salvage work in advance of the illing of the Ilısu Lake has been on-going annually since 1997. he broad project goals established during the irst ield season continue to guide our research project: (1) to study city planning and urban functioning at a major Assyrian city along the imperial frontier; (2) to document the entire occupational sequence at Ziyaret Tepe and within the upper Tigris River valley; (3) to explore the relationship between the Assyrian and indigenous Anatolian populations; and (4) to study the impact of urbanization on the regional Late Bronze and Iron Age ecology and environment. In the operation reports that follow, we situate the speciic ield objectives of the excavation seasons within this broader research agenda. the British Institute for the Study of Iraq, a number of private donations paid through the Ziyaret Tepe Archaeological Trust (UK), and the University of Akron. Finally, we would like to acknowledge the long-lasting and important support of the editor of Anatolica, Dr. Jacob Roodenberg, who has gracefully allowed us access to this vital forum for presenting our work. Anatolica XLI, 2015 3 he site of Ziyaret Tepe can be divided into two main morphological units: a high citadel mound with an area of approximately 3 hectares rising approximately 22m above the surrounding agricultural ields and an adjacent lower town, occupying approximately 29 hectares to the east, south, and west of the citadel mound (Fig. 2). During the 2011-2013 seasons, major excavations were undertaken on the high mound, in Operation N, a monumental mudbrick building which we have nicknamed the Bronze Palace. Eight excavation areas in the lower town were also excavated, including ive new excavation areas in the southwestern lower town in Operations T, U, V, Y, and Z and the expansion of previous excavation areas in Operations K (2013), M (2012) and G/R. his latter expansion was excavated as Operation W. In addition to excavation, we also undertook three additional ield seasons of subsurface geophysical prospection in the lower town using electrical resistivity survey; the geophysical prospection is not covered in this report. Ziyaret Tepe had a long occupational history spanning the Early Bronze through Middle Iron Ages, with periodic later occupational levels in the Late Iron/Hellenistic, Late Roman, Medieval, and Ottoman periods (Fig. 3). he zenith of occupation at Ziyaret Tepe was as an urban center in the 9th through 7th centuries BC during the Late Assyrian period. Historical inscriptions explain that the city of Tušhan was re-founded following a hiatus brought about by the Middle Assyrian political collapse of the mid-11th century BC. In 882 BC, King Ashurnasirpal II claimed to have rebuilt the city which was continuously occupied by the Assyrians until its abandonment in 611 BC after the collapse of the imperial heartland (Grayson 1991; Roaf in Matney et al. 2002: 49-51). hese Assyrian historical accounts correlate closely with the architectural history and stratigraphy of the excavated remains at Ziyaret Tepe. Excavations in Operation A/N: The “Bronze Palace” (D. Wicke) Since its initial discovery in 2000, our primary interest in Operation A/N has been to untangle the complex ground plan and chronology of an Assyrian palace which we refer to as the “Bronze Palace” (Matney et al. 2011: 69-72). he principal objectives in Operation A/N in the 2011-13 seasons were to amend and expand the architectural plan of the palace to the north where Assyrian walls appeared immediately beneath the surface and to investigate earlier stratigraphic levels irst seen in the previously-excavated ‘Deep Sounding’ excavated in Room 9. It is now clear that we have three separate Assyrian building phases in the Bronze Palace, here labeled I, II, and III. Phase I corresponds to the uppermost phase, probably 7th century BC. Phase II refers to the main building phase of the mid-8th century BC date, which was destroyed in a heavy conlagration. Phase III shows up immediately beneath phase II-walls/foundation and might, in fact, have been the 9th century building mentioned by Ashurnasirpal II (Grayson 1991: 202), although this needs conirmation by continued analysis of the small inds. In 2012, we discovered possible evidence for an “intermediate” stratum of possible Hellenistic to Late Antique date, which consists of a loor of reused baked bricks associated with a trefoil-mouthed jug (ZT 41280) and a pebble-surface in grid square N1010E1170 and possibly a pisé wall in N1010E1190. In 2013 more evidence for a pre-Late Assyrian level became clear. 4 Timothy Matney et al. Fig. 2. Topographic plan of Ziyaret Tepe showing the areas of excavations, 1997-2014. Fig. 3. Chronology of the operations. Anatolica XLI, 2015 5 Bronze Palace: Phase I Excavations in 2011 brought about some additions to the plan in the northern areas through the excavations of grid squares N1010E1160, N1010E1170-1190, and a 2.5m strip across N1020-1022.5E1170-1190. Especially informative was the excavation of Room I.13, attributed to the uppermost Phase I. his room has a loor of baked bricks, covered in bitumen; very small pebbles seem to have accumulated in the corners, trodden into the bitumen. Along the edge of the rooms, bricks are standing on edge; a sink-hole in the south wall marks the beginning of drain (N-748), running south. he drain, however, was found to be destroyed and apparently went out of use. he pavement extended over the actual wall-line in the SE-corner, where there was a doorway to the southern Room I.15. Two possible doors can be reconstructed to the NE, where three bricks extended over the bitumen line which elsewhere indicated the ancient wall line, as well as to the NW, where a pit destroyed the wall. he north wall is attested by a few brick lines only due to is being very close to the surface. Our excavations suggest that Room I.13 was a remodeling of a phase II room, following its extent and reusing its walls, but using diferently colored bricks, which were laid less well. Drain N-748, apparently, was not repaired for that later reuse. In N1010E1170, vessel ZT 38079, protruding from the modern tell surface, turned out to be a tall lask, one of the grave goods associated with Late Assyrian burial N-581. his burial was disturbed by a later tannur (N-575, probably of Medieval date) which seems to have cut away the lower part of the body as its legs are missing completely. Grave goods included a ibula (ZT 38130) indicating a date in the 7th century BC or later according to Pedde’s typology of ibulae (Pedde 2001). his date is in agreement with its stratigraphic position above the loor to Room II.11 and, therefore, postdating Phase II. Bronze Palace: Phase II he recent excavations also clariied a number of important details relating to the 8th century BC palace (Fig. 4), despite extensive pitting of Medieval date. One principal observation was in grid square N980E1160 which was excavated to examine a major Fig. 4. Plan of the Phase II Bronze Palace, Operation A/N. 6 Timothy Matney et al. north-south running mudbrick wall (N-510) that appeared to form the western wall to Room 7. It was established that wall N-510 turns east at its southern end and does not continue into the baulk. his is signiicant because from this observation we can conclude that Room 7 immediately gave way into the western courtyard of the palace complex. Our expectation, given the “standard” form of Assyrian palaces, was that there would be an antechamber west of Room 7, but this appears not to have been the case with the Bronze Palace. Also in N990E1170, south of pit N-651, we found a cross-wall demonstrating that the building plan had an additional room, Room II.14, located to the west of Room II.7a, and having the efect of reducing the size of the courtyard. A number of other additions and modiications can now be made to the Phase II plan of the palace. (1) Our excavations suggest that Room II.13 was a remodeling of an earlier room. Following the 2010 excavations it was clear that Room II.3 had a doorway in the northwest leading to an unexcavated room. Room II.3 leads opens into Room II.11 and was therewith connected to Room 13. A southwest annex to Room II.11 appears to be a well or sink. Only the uppermost area could be investigated and appeared as a built construction, similar to a large cess pit N-511 beneath Room I.9. Room II.11 posits some problems, for no proper loor was found to it. Room II.13 was connected to Room II.15, which was equipped with a small bench and canal-mouth in its NE-corner. A major drain (N-735+741), constructed out of baked-bricks and stone, was running east to west through the room. A connection to the large cess-pit N-511 at the N-corner of Room I.9 appears to be quite likely, but cannot be proven due to two large pits, which cut into the southern part of Room II.15 and its south wall. Room II.15 gave access to Room II.17, a narrow dead-end of about 80cm in width, which might be interpreted as a staircase. To the northeast, the existence of Rooms II.12 and II.18 could be proven by scraping the surface; their attribution to Phase II was made because of their orientation and the use of similar bricks as in Room II.4/8 and II.1. A re-investigation of Room II.1 conirmed the hypothesis, that it was constructed as a separate antechamber in Phase II. Its function changed to an open alcove, however, after a remodeling to Phase I: the southern wall was clearly laid over by courtyard-bricks. In 2012, a deep sounding showed evidence that the eastern part of the Phase II building required a mudbrick foundation platform because there was a pre-existing erosion gully which created uneven terrain for the 8th century BC construction. In 2013, work concentrated in trenches N1000 E1170 and N1000 E1160, where a deep sounding was irst opened in 2009. he west section of this excavation area had eroded heavily during the last few years; therefore, it was cut back by 2m to the west along its full length with a goal of adding to our plan of the building and clarifying the nature of the rooms that lay to the west of the previous excavations. In the northern part of that western extension of the deep sounding, it was possible to trace the remains of the Phase II walls delimiting Room II.19. he collapse of the Room II.19 included fragments of painted wall plaster, similar to those described in earlier reports from Room II.7, but in poor condition. Immediately below, an earlier building phase (Phase III) followed the same basic plan supporting the conclusion that the Phase II palace immediately followed the Phase III palace chronologically, using its old walls as foundation. Anatolica XLI, 2015 7 Bronze Palace: Phase III When clearing the disturbed area along the western section in Room II.9, a well preserved brick pavement (N-630) came to light, thickly covered in bitumen with an associated wall N-671. he bricks of Wall N-671 were unusual; they were large (up to 90cm by 60cm) and were made out of coarse reddish clay containing a fair amount of limestone, set in grey mortar, with unusually thick mortar-lines. N-671 forms one side of a room, at least 3.60 m in length, which, signiicantly, precedes the Phase II building. he wall could not be excavated in its full width, since the outer faces are hidden, respectively beneath walls N-112 and N-265. Wall N-671, as well as pavement N-630, is set atop an inill of baked brick fragments, rubble, and large sherds in a reddish clayish matrix. It appears possible that Wall N-671 and pavement N-630 represent vestiges from a 9th century BC version of the Bronze Palace. By excavating through the palace loors, we were able to observe subloor construction techniques used by the Assyrian builders. As documented earlier, the Bronze Palace had an extensive internal drainage system. In 2013, we exposed canal N-835 that led into cess-pit N-865 (Fig. 5). Incorporated in the brick lining of the canal was half a brick with a gameboard scratched into it (ZT 41642). he canal was cut by pit N-811, and a baked brick with semicircular hole in a direct line to the east running through wall N-265 indicates its point of origin. he canal and cess-pit would have served Room III.6. Another drainage system, excavated previously in this area, serviced Room III.9. he cess-pit was completely preserved with false vaulting in situ but was only reached in the inal week of excavation. he uppermost six courses of the cesspit are mudbrick. he sevFig. 5. Canal N-835 exposed in Operation enth course from the top mixes limestone A/N which served as part of the drainage system inside the Bronze Palace. blocks and bricks. From the eighth course downward the sides of the cess-pit are made of unworked or roughly hewn limestone blocks of small to medium size. he bricks and stones are covered with a whitish layer of salt or phytoliths. he pit itself was empty for the top 1.40m and comprised rather soft, blackish-greyish material with a few bones and brick fragments. Summary: Operation A/N Excavations he earliest occupation remains discovered at the bottom of Room 6 did not reveal any clear architectural features. Wall N-539 and stone paving N-363 are to be dated to the Middle Assyrian period based on stratigraphy, a preliminary reading of the pottery, and the 8 Timothy Matney et al. unusual thick (ca. 12 cm) bright orange bricks laid in light-grey mortar comprising N-539. However, the construction of stone wall foundation N-839 is diferent from the typical Middle Assyrian wall foundations and Dr. Wicke has argued on stratigraphic grounds that N-839 must represent a pre-middle Assyrian phase and, accordingly, that the deposits of N-854, N-857 and N-858 underneath ought to be considered as Late Bronze Age or earlier. In clearing Rooms 6 and 2, work was considerably hampered by the presence of two large pits. One of these pits (N829) contained an unbaked clay tablet (ZT 41575) mixed with brick rubble that was thrown into the pit while building up the foundation to that room. he tablet, to be published fully in a subsequent publication, is early Late Assyrian in date and lists 25 cloth garments. In sum, before the Late Assyrian building was constructed, there must have existed a Middle Assyrian predecessor building incorporating a stone-paved courtyard and considerable mudbrick walls. Lower Town Excavations Eight excavation areas were excavated during the 2011-2013 ield seasons under the general direction of Dr. John MacGinnis. Two of these were horizontal expansions of earlier excavations; both Operations K and M were undertaken with the speciic goal of recovering more of the plan of domestic buildings recovered earlier and to secure additional well stratiied archaeobotanical and zooarchaeological samples for analysis of commoner dietary and economic practices. As detailed below, the excavation of three soundings in Operation W were undertaken to answer speciic questions about the monumental mudbrick building in the lower town previously excavated as Operation GR. Additionally, ive new operations commenced with the goal of investigating speciic structures targeted after geophysical prospection. he results from these excavations both allow us to now reine our interpretation of the plan of Assyrian Tušhan, and to document a previously unrecognized occupation of the site in the Late Antique period (5th-6th centuries AD) in Operations T and U. Operation K: Assyrian private housing Operation K is located in the southern lower town, in an area associated with the Late Assyrian city wall. he irst excavations in Operation K were conducted under the supervision of Dr. John MacGinnis in 2003 and Prof. Dr. Kemalettin Köroğlu in 2004. During these previous two seasons, an area of 20m by 10m was excavated, resulting in the discovery of the city wall running in a SE-NW direction, and a domestic residence constructed immediately inside the city wall. he city wall was constructed of mudbrick on top of an artiicially raised terrace. he area outside of the wall was deeply trenched, creating a large moat in order to increase the city’s defenses as well as to possibly direct water coming from a small tributary to the south of Ziyaret Tepe through channels situated outside the walls and ultimately draining into the Tigris River. he 2013 work in Operation K aimed at exposing the eastward continuation of the Late Assyrian domestic building discovered during the 2003-2004 excavation seasons, and to determine its phases of use, complete plan, and function. Anatolica XLI, 2015 9 Previous excavation in Operation K had unearthed a building comprising four rooms (A-D), a courtyard (E) and a semiopen space to the north of the courtyard, dated to the 8th-7th centuries BC. It is evident from in situ tannurs that two rooms (A and E) were used for cooking purposes. he courtyard to the north was paved with pebbles and re-used bricks, with a deep well (K-098) in the middle. he mudbrick walls were 80-90cm thick and excavations in 20032004 showed two phases of occupation. New excavations in 2013 were conFig. 6. Operation K domestic architecture at Ziyaret Tepe. ducted to the east of the area dug in 2003-2004 in trench N750 E920. In 2013, the eastern half of Room C and two additional rooms were discovered. he plan of the Operation K building now comprises eight rooms (Fig. 6). Five of these rooms (B, A, C, G and H) are situated on a NW-SE axis, adjacent to one another and parallel to the city wall for the course of about 20m. he newly excavated part of the building in Operation K also had two occupational phases and it was conirmed that the same ground plan was repeated in both phases. he mudbrick foundations of the later phase walls are somewhat better preserved than those of the earlier phase. Individual mudbrick sizes of both phases are similar, ranging between 38cm and 40cm on a side. he walls comprise alternating courses of two full-bricks followed by a course of one half-brick set to either side of a full-brick, yielding wall thicknesses are of 80-90cm, identical to those discovered in 2003-2004. Room C was 3m by 3m in extent and has loors and associated artifacts corresponding to both phases. he northern and eastern walls of the earlier phase were discovered at foundation level. he most important unit of this phase is a simple earthen grave dug into the mud loor aligned in a north-south direction. he head of the skeleton inside the grave is on the north, with the skull facing eastwards. he right hand was placed on the stomach, and the left hand was resting to the left side of the face. In the later phase (K-125), the loor was raised by approximately 20cm and a hearth (K-151) was placed above the earlier phase grave. Immediately beneath this later phase loor, there are two burials running parallel to the eastern (grave K-117) and southern (grave K-119) walls. here are no inds in the graves except a vessel base in grave K-119, where only the foot bones of the body are preserved. he entrance to Room C is reconstructed to be in the north. A possible baked brick threshold of the earlier phase and the 10 Timothy Matney et al. pivot hole/door socket of the door were discovered on the northern wall in the northeast corner of the room. he later phase door must be on wall K-080 that runs in the same direction. he northern walls of the two rooms east of Room C were damaged by numerous later pits. In some cases the walls can only be traced below the foundation loor level on the loors of the pits. Some mudbrick remains of a wall dividing rooms G and H were discovered at the bottom of a later phase pit. Room G is 3.0m by 2.5m in extent. Of the earlier phase walls of this room, the southern wall is discernable, whereas the other walls can only be traced as foundations. here are no in situ inds on the loor of the earlier phase. he western (K-111), northern (K-136) and southern walls (K-110) of the later phase are discernable. he loor of this phase is not preserved, most likely due to its proximity to the surface. Room H was poorly preserved. he mud loor (K-129) of the earlier phase of Room H was exposed. However, the later phase loor could not be identiied. he northern and southern walls of the earlier phase of this room are discernable, whereas only traces of the eastern and western could be revealed. In the later phase, all walls (K-127, K-132, K-152) but the western wall were delimited. Fig. 7. Photograph of Late Assyrian building in Operation K facing west. he meterstick sits on the lower loor of Room G. Room C extends into the far baulk. he damage caused by the later phase buildings and pits is more extensive to the north side of the Late Assyrian building. he fact that the entrance to Room C and Room E is in this area and that traces were uncovered of a possible paved threshold in wall K-122 between Courtyard E and the open area suggest that the area to the north of rooms G and H was used as a general purpose space. In 2013, in addition to Late Assyrian pottery, the inds of two ibula fragments, one stone vessel fragment, one stamp seal and two beads aided the dating of this building in Operation K. It is notable that this material is of a generally lower quality and Anatolica XLI, 2015 11 quantity than found at other more substantial domestic residences across Ziyaret Tepe, suggesting that this was an area of commoner housing. In Operation K, on the northern side of the trench N750 E920, remains of a completely diferent character than typical Late Assyrian architecture were discovered just below the surface. he poorly preserved architecture of this occupational level consists of two drainage pipes and associated rows of oval paving stones, and groups of stones that could be post holes or bases of some sort. Similar rows of stones were discovered in previous excavation seasons both in Operations K and T and are the subject on on-going analysis. Summary: Operation K excavations Work in Operation K in 2013 allowed for the clariication of some previously indistinct details and correction of various earlier assessments. For instance, it is now understood that the courses of mudbrick which we considered to be benches in Room A and Room C were the walls of the earlier phase (cf. Matney and Rainville 2005: 33-35; Fig. 10). It is also understood that after the destruction of the earlier phase building, the rubble was cleared to the foundation level and the earlier phase loors were raised by 15-20cm. In the later phase, the building was shifted 80-90cm southwards and was rebuilt using the same plan. here is no evidence of ire or destruction in the Late Assyrian buildings. here is also continuity in use through both phases. Room C, excavated in 2013, appears to have been chosen for intramural burial during both phases. his situation is similar to Room A and Room E, excavated in 2003-2004, being used as possible kitchens with tannurs in both phases. Consequently, it can be argued that there is not a long occupational gap between the two phases. Zooarchaeological report Operation K, 2003-2004 (T. Greenfield) his report describes the zooarchaeological remains collected during the excavation of the Late Assyrian domestic structures in Operation K during the 2003 and 2004 ield seasons by Prof. Kemalettin Köroğlu. he faunal remains from the 2013 excavation season are still under analysis and are not included here. As such, the discussion is limited to those remains uncovered by the end of the 2004 season: four rooms (A, B, C, D) and a central stone courtyard (Room E) located to the northeast of those rooms (see Matney 2005: 32-35 for an initial description). As noted above, based on location of the structure, and the nature of associated material remains and architecture of the building, the excavators argue that these rooms housed commoners living just inside the city’s fortiication wall. Sampling protocols he entire sample of faunal remains was collected according to site-wide protocols. All primary contexts (supraloor or just above, deined pits/features) were either sampled and sieved on a 1:5 bucket ratio or fully sieved depending on the feature excavated. Secondary contexts (ills and building collapse) were hand-collected as part of the larger excavation process 12 Timothy Matney et al. (Greenield in Matney 2009). Micro-fauna which were collected as part of the lotation or micro-debris samples are not included in this report and this therefore can lead to a bias in the report towards larger animals. Taxonomic frequency he total sample of bones for the building was 468 pre-excavation fragments. No articulations were present in this sample and each fragment appears to be from a separate bone element and/or individual. Only mammals are present in this sample. Initially the sample included two human specimens which have been removed from the analysis, which decreases the total to 466 fragments. Completely unidentiiable fragments were counted (n= 9) but not used in the inal analysis of taxonomic frequencies. his brings the sample number down to n=457. Identiied to a genus or more speciic level there are 270 specimens when the general categories of mammal (large, medium, and small) are excluded. he 270 fully identiied specimens will be discussed in relation to state of domestication and element representation. Domestic versus wild taxa he overall animal assemblage is overwhelmingly composed of domestic taxa. his implies exploitation and animal husbandry strategies are primarily geared towards the consumption of domestic animals (sheep/goat, cow, and domestic pig). Domesticated animals dominate with 92% followed by wild specimens at 8% of the total sample. Domestic taxa: he highest frequency of remains is from Caprines (Ovis aries (sheep) and Capra hircus (domesticated goats) (n= 154) (Fig. 8). Of these, most elements can not be distinguished between Ovis and Capra (n= 106). hirty-one elements are distinctly of Ovis aries while 17 are from Capra hircus (n= 17). Bos taurus (cow) is the second highest taxa frequency with (n=40), followed by Sus scrofa dom. (pig) (n=35). here is a drastic decline in the frequency for the remaining domestic animals. Canis familiaris (domestic dog) has the next highest frequency (n=5), followed by equids [E. asinus (ass) and E. caballus (horse) with 2 specimens each]. Wild taxa: he highest frequency comes from Lepus sp. (hare), (n=7), followed by Gazella gazella as the most frequent species (n=4), Dama dama sp. (fallow deer) with three specimens, and Capreolus capreolus (roe deer) with two specimens each. Cervus elaphus (Red deer), Anura sp. (frog), Rodentia sp. (rodent) and Sus scrofa fer. (boar) each are represented by a single element in the sample. hese species distributions are starkly diferent from what was found in the “Bronze Palace” on the upper citadel at Ziyaret (Greenield in Matney 2011). For example, turtles/tortoises and exotic birds are found only in the palace. Element distribution One indicator of status or wealth in ancient complex societies is linked directly to what types of meat were consumed (Curet and Pestle 2010; Grant 2002). Analyses of body Anatolica XLI, 2015 A. Operation K % frequency of wild taxa (n=20) 13 B. Operation K % frequency of domestic taxa (n=238) Fig. 8. Pie chart showing the distribution of domestic and wild species represented in the Operation K zooarchaeological assemblage. elements or portions of the body that are consumed is an integral part of the process in determining social and economic factors within a settlement (DeFrance 2009; Marom et al. 2009, Reitz, 1987: 109). A brief description of portion preference of animals from Operation K is provided below. In terms of domestic animals utilized in the Operation K structures, there is evidence that all body portions (cranium, proximal limbs, thorax and distal limbs) of the sheep and cows were consumed including the undesirable ones with little or no meat (i.e., distal ends of limbs). here is no indication of a preference for speciic cuts/portions which would have yielded the highest meat weights (e.g., upper anterior and posterior limbs). he entire carcass was used for most of the domestic species. he exception to this pattern is for Sus scrofa dom. which had a clear preference for cranium and thorax (ribs) elements over other parts of the animal. In terms of wild animals, the majority of the sample is comprised of smaller animals that could easily be eaten as a whole unit without portioning them. he larger animals (Cervus elaphus and Capreolus caprelous) are the only two samples used to investigate element/portion preference and even these numbers are really too small to give an accurate picture of portioning. he data however, did show there was a small increase in the posterior and thorax portions over the rest of the body in both of these species indicating a slight preference for this part of the animal of the body. Again however, these selected portions of the body are and are not usually indicators of elite status and these results suggest the utilization of only parts of the carcass in the domestics but must be used with caution due to the small sample size. Analysis of element preference for several domestic species (cow, sheep/goat and pig) and two wild species (red and roe deer) indicates a clear pattern of consumption. Most or all parts of the animal remains from in the house were consumed equally with a very slight prefer- 14 Timothy Matney et al. ence for posterior portions in wild animals. Element preference for both categories of domestic and wild taxa does not seem to be the driving force of food consumption, but rather a case of utilitarian means and access to any available meat. his pattern of consumption is indicative of low status households (Greenield 2014). he pattern of high frequency of low status wild animals (rabbit, small deer, rodent and wild pig) in the house conirm the occupants as lower-ranked members of society. Diets are sometimes supplemented in the lower status houses, possibly as part of a diversiied subsistence strategy (Reitz 1987). he combination of the presence of elite species and a clear element preference from the palace gives an indication that diet was somewhat geared towards status and prestige in this area. his is clearly not the case for Operation K. It was not until the both faunal samples were fully analyzed that the diferential patterns emerged between the two areas. Summary: Operation K Zooarchaeology he faunal remains from this building indicate the use of wild and domestic animals for utilitarian purposes. Based on these preliminary results from Operation K it is evident there is a predominance of domestic fauna with a supplement of low status wild animals into their diet which difers signiicantly from the pattern seen in the elite areas of the site such as the Bronze Palace. While there is a preponderance of the traditional domestic species (cow, sheep/ goat and pig) in this house also endemic to most of Mesopotamia at this time, the interesting details are with the distribution of wild fauna at Ziyaret Tepe. Operation M Operation M is located to the south of the mound, where the lower city is situated. Here the excavations were conducted by Prof. Kemalettin Köroğlu and Harun Danışmaz in 2012. he trench N760E790 dug here measures 10m by 5m. A previous trench opened in 2004 measured 1.5m by 10m and revealed architectural remains belonging to two distinct building levels, and the presence of well-constructed street (Matney and Rainville 2005). he 2012, work brought two building levels to light. he latter level (M1) relects the characteristics of small-scale occupation, while the lower level (M2) comprised substantial mudbrick architecture of the Late Assyrian period. hese results parallel those of the previous excavations. Level M1: Pit Level Beneath the 30cm deep plow zone, possible faint traces of the irst building level were discovered. A pebble and sand layer in the south of the trench and fourteen pits belongs to this level. Pit M-103 in the southwest of the trench yielded glass and pottery fragments. A nearly complete pot (ZT 42014), two earrings (ZT 420009, ZT 42025) and a ring (ZT42012) found in the north of the trench, immediately beneath the agricultural soil might be related to this level though they were not in situ. Dating of this strata is uncertain. Anatolica XLI, 2015 15 Fig. 9. Late Assyrian building in Operation M, Level M2. Level M2: Late Assyrian Building Traces of melted mudbrick walls were encountered 60cm beneath the surface, but remains of what can safely assumed to be walls were found at a depth of 1.0m. he architecture of this level comprises two rooms and remains of a street (M-128) in the south of the trench (Fig. 9). Room A (M-133) measures 3.4m by 5.0m. Room B was found in the northwest corner. he walls are between 1.0m to 1.1m in thickness, built with courses of two and a half bricks placed side-by-side. he walls do not have stone foundations. It was determined that this structure had two phases and was dated to the Late Assyrian period through small inds. 16 Timothy Matney et al. In the early phase, foundations of the north (M-110), west (M-109) and east (M-130) walls of Room A have been excavated. A door pivot stone found 1.1m north of the southeast corner was placed here in the later phase, which might point to the fact that the entrance was at the same spot in the early phase. he foundation trench of the south wall runs parallel to the pebble-paved street, suggesting that the room and the street (M-128) were contemporary. he interior face of the wall (north) is not clearly observable. he extant foundation trench shows the possible existence of a thicker wall or a bench adjacent to it. he loor of Room A was made of pressed mud and produced no inds. In the center of the loor is a pit 50cm in diameter and 1.0m in depth (M-135). A similar pit was discovered in Room B to the northeast. he most signiicant inds of the early phase of Level M2 were unearthed in two graves: M-131 and M-134. Grave M-131 his grave was dug in the north corner of N760/E970 trench, into loor M-136. It measures 1.00m by 0.80m and has two child burials. he later burial was placed in the north of the grave, 20cm below the room loor, while the earlier one was in the south and 45cm below the loor. he skeletal remains of both children are not well preserved. Of the skeleton on the top only pieces of skull, ribs, arm and vertebrae were uncovered. he only ind from the burial is a bronze bracelet on the left arm of the higher skeleton (ZT 42151). he lower burial’s bones are fragmented except for the skull. A bowl (ZT 42153) was placed below the rib cage. Grave M-134 his grave was dug at the edge of wall M-109, into the loor M-133 (Fig. 10). he cut of the grave measures 0.70m by 1.85m and it was extended as the pit deepened. he skeleton was initially measured at 1.70m tall and was laid face down. he right hand is near the right shoulder, while the left hand is below the chest. All the grave goods were found near the skull Fig. 10. Grave M-134. Anatolica XLI, 2015 17 and ribcage. he function and the dating of the building were determined in the light of the grave inds. he bowls in both graves, and beakers, ibulae and beads are similar to those from the Late Assyrian period. he ibulae in particular are dated to the 8th century BC, providing a rough date for this burial. In the late phase, the plan of the room related to loor M-133 was preserved, but there are indications that its function changed. A new loor (M-124) was formed by raising the earlier one. A tannur (M-116) discovered in the south of the room belongs to this phase. he entrance is not from the street to the south, but from the east wall. A door pivot stone found on the east wall related to a stone-paved entrance immediately to the west. he grave goods indicate that the earlier structure belonged to wealthier individuals. Modiications in the later phase, which included a hearth, suggested a rearrangement of the earlier spatial layout of the room. Operation T Operation T was a new excavation area in the southern lower town at Ziyaret Tepe. Two 10m by 10m grid squares, N770E1070 and N760E1070, were chosen for excavation as the result of a previous magnetic gradiometry survey in 1999 (Matney and Bauer 2000) which suggested that Operation T represented an area of private housing bounded by streets on the northern and southern sides. he aim of the operation was, therefore, to explore the domestic housing of the ancient city in order to complement exposures of large public and elite housing excavated elsewhere. Work in Operation T was directed by Prof Dr. Kemalettin Köroğlu. Level T1 he uppermost occupation levels in Operation T comprised imperfectly preserved stone wall foundations with associated cobbled surfaces in both trenches. No good loor surfaces were found for this level but there were a number of associated tannurs. he inds in these upper levels included many pieces of broken roof tiles, ceramics including strap-handled jar fragments and glass fragments, including a glass funnel (ZT 39068). he area was heavily pitted, of which two pits (T-005 and T-009) had roof tile fragments on the surface, one pit (T-007) was illed with a heavy concentration of broken roof tiles and yet another pit (T-017) contained a single complete tile. Many of the other pits contained either ashy ill or a soft earth ill suggesting that they held domestic waste. Level T2 Underlying these remains was an earlier phase (T2) which was much better preserved. he architecture of this lower phase consisted of foundations made of larger stones than in Phase T1, together with an intact loor (T-031). (Fig. 11) On the loor lay a well-preserved domestic assemblage including a large terracotta tub (ZT 39133), three small bowls (ZT 39188, ZT 39189, ZT 39206), two small lasks (ZT 39315, ZT 3930), and a basalt potstand (ZT 18 Timothy Matney et al. 39264). Of particular interest was a sherd from a pilgrim lask with an elaborate incised decorated (ZT 37204). he small inds included a bronze ring (ZT 39073) and a number of objects made of iron: a ring (ZT 39134), two keys (ZT 39067, ZT 39104) and two iron plaques, and perhaps horse ittings (ZT 39314). A large lat stone, perhaps for grinding grain, also lay on the loor. Particularly Fig. 11. Late Assyrian building in Operation T, Level T2. important for dating was Floor T-031 inside stone walls foundations. Photograph a coin of Justinian found of northern part of the operation, from the west. on the loor and a coin of Constantine II found below the loor; a third coin was also found in the baulk. he second of these gives us a terminus post quem for the abandonment of the building. In an attempt to expedite the discovery of remains of Assyrian date, the southern square was divided and a 5m wide section was delineated across the northern half of that square for deeper excavation. his trench was excavated to 30cm below the surface of the Phase T2 remains but no Assyrian remains were recovered. he area continued to be alicted by heavy pitting. After these pits had been excavated to depth of a further 50cm without any solid evidence for Assyrian remains, the attempt was abandoned. At this stage, our evidence suggest that the interpretation of the magnetic anomaly running obliquely across the area of Operation T was an Assyrian street is either incorrect, or that this feature is buried more deeply than anticipated. he coins Initial identiication of the coins in the ield was undertaken using on-line catalogues examined for comparanda. We are grateful to Edward Dandrow for corrections and for further information on these coins. (Fig. 12) Coin ZT 39223, found in situ on the Phase T2 loor, is a copper alloy coin dating to the reign of Justinian (AD 527-565). he obverse shows the emperor’s head in proile wearing a ribbon or laurel leaves with the inscription DV IVSTINI-ANVS PP AVG around the edge; the reverse has a large M with three dots underneath, a star on the right, a cross on the left, and another cross above. It was minted in Constantinople. Coin ZT 39240, found in a tertiary context, is a copper alloy coin dating to the reign of Arcadius (Flavius Arcadius Augustus), emperor AD 383-408. However the dating of this particular coin can be further narrowed as this reverse was only in use when he was Emperor in the east, AD 395-401. he obverse depicts the emperor himself while the reverse shows the emperor Anatolica XLI, 2015 19 holding a sceptre and shield with the igure of Victory standing to his left crowning him and holding a wreath above the emperor’s head. he obverse carries the inscription: DN ARCADIVS PF AVG. he reverse inscription reads: VIR[TVS EXE]RCITI. his coin was minted in Cyzicus. Finally, coin ZT 39345, found under the Phase T2 loor, is a copper alloy coin of Constantius II, emperor AD 337-361. he obverse has the head of the emperor with a ribbon in his hair and an inscription efaced but perhaps to be read: [DN] CONSTAN-TIVS PF [AVG]. On the reverse the motif is a Roman soldier spearing a Persian horsemen; it is however possible to partly make out the inscription, perhaps FEL TEMP REP[ARATIO]R. his coin was minted in Constantinople. Operation U Fig. 12. Roman coins from Operation T. a: ZT 39233 (Justinian); b: ZT 39240 (Arcadius); c: ZT 39345 (Constantius II). Investigations were carried out in Operation U, situated in the far southeast corner of the lower town, in the 2011 season. We were keen to work in this locality as it was a part of the site where we had not hitherto conducted any excavations, but our interest was particularly focused by the results of the mapping by magnetometry carried out in 2009 (Matney et al. 2011: 94, 112, Fig. 18) which had produced indications that there might be a series of three substantial residences situated in this area close up to the city wall. However, intriguing as these images were, they were faint and in need of corroboration. Accordingly, a main aim of the work in Operation U was to further investigate these features and we approached this by a combined plan of resurveying the area using resistivity followed up with targeted excavation. Operation U Geophysics he geophysical survey was carried out as follows: having secured the permission of the land owner, a 40 x 40 m area was staked out (with the southwestern corner at N720E1340) and the stubble of the wheat crop, which had been harvested, burnt of. his 1600 m2 area was then subdivided into a grid of 10 by 10m squares and in the course of the season the entire area was mapped using electrical resistivity. In the course of preparing the area for the geophysics a Paleolithic handaxe (ZT 40441) was found in the southeastern square of the operation (N720E1370). he electrical resistivity survey generated a signiicantly more detailed and comprehensible ground plan of the building than previous magnetic gradiometry (Fig. 13). 20 Timothy Matney et al. Excavations in Operation U In order to ground truth these results and to gain an insight into the stratigraphy of the area we also conducted a targeted excavation in Operation U. he grid square selected for excavation was N730E1340, chosen because it appeared to straddle the western side of the building and might therefore be expected to give an idea of both internal and external areas. To our surprise, the irst intact remains we came upon were Late Roman in date. Starting in the plough zone Fig. 13. Preliminary presentation of processed electrical resistivity data from Operation U. a large quantity of Roman pottery was found as well as a few fragments of glass and an iron nail (ZT 40105). In the end, we recovered two Roman layers, overlying Assyrian levels of which direct evidence for two phases was recovered and with every indication that further phases might lie below. Cutting into all of this were numerous pits, the contents of which were Late Roman in date. One pit, U-069, contained a well preserved limestone roof roller (ZT 40277), 5 cm long with a diameter of 21cm and weighing 48kg, evidently thrown into the pit, along with other stones, in order to create a solid foundation for a subsequent building level. A broken rock crystal bead (ZT 40059) was found in pit U-016 and an iron bell (ZT 40272) in pit U-071. Phase U1: Upper Roman phase he upper Roman phase was characterized by stone wall architecture. his was only fragmentarily preserved. he main wall of this phase was a stone wall measuring 0.75m wide, consisting of a facing of stones on the outer sides and a mud inill, most probably foundations. he wall was cut in two places and consequently consisted of three stretches, U-002, U-005 and U-006. he surfaces associated with this phase were U-003 to the south and U-004 to the north. he former of these was not a well laid surface but rather a horizon comprising a thick layer pebbles and probably representing an outside surface. U-004 was similar but with some fragments of mud plaster; it would therefore appear that this was an interior area. Anatolica XLI, 2015 21 Phase U2: Lower Roman Phase Removal of these remains exposed a second Roman layer of which the most striking aspect was that the walls were built out of mudbrick. he later phase was built almost directly over the earlier one and with little or no intervening space of time. Two walls were identiied in this phase, a northerly wall (U-023) and a southerly wall parallel to this at an interval of 2.60m (U-031). Wall U-031 was a brick and a half wide, the full size bricks measuring 48 by 48cm square (thickness unknown). hese are very large dimensions for mudbricks - considerably bigger than the standard Assyrian size. Wall U-031 was one and a half bricks wide, with a total width of 75cm. In the case of the more northerly wall U-023, it was extremely diicult to articulate the individual bricks or indeed to discern what the standard brick size in this wall was: possibly half bricks measuring 20cm by 40cm but there appeared to be many irregular pieces. A quantity of Roman ceramics and glass continued to be associated with this phase although far fewer roof tile fragments. his may suggest that the roof tile fragments found in the upper phase came from the roof of the lower phase building. he surfaces associated with this phase were cobbled surfaces. here was one in situ door socket installation comprising a white limestone door socket (ZT 40152) irmly emplaced on a broken stone trough (ZT 40179) stood on end over two stones with a large sherd wedged underneath. Phase U3: Upper Assyrian Phase With the removal of the lower Roman phase we came down on to the remains of a Late Assyrian building (Fig. 14). he irst indication of this was a mudbrick wall running approximately southwest-northeast in the eastern half of the trench (U-040), with a wall coming of at right angles to this in the southeastern corner (U-043). In due course further walls were identiied: U-085 coming of U-043 to the south, U-060 coming of U-040 to the east, and another return of the main wall, U-077, in the northeastern corner of the trench. his layout of walls delineated three main spaces, Room A in the southern part of the trench in the angle south of U-040 and U-043, Room B east of U-040 and between U-060 and U-066, and Room C east of U-040 and between U-066 and U-077. After removal of the upper Assyrian phase the term Room B’ was used to denote the whole space east of U-040 and between walls U-043 and U-077. he walls of this building were made of pink mudbrick measuring 40 by 40cm on a side with a thickness of 8-9 cm. Both the size and the use of the local clay are typically Assyrian. he areas exposed were excavated down to loor level. Fig. 14. Late Assyrian building in Operation U, Level U3. 22 Timothy Matney et al. Although there were no exceptional discoveries in terms of small inds, all these areas yielded excellent environmental material in the form of zooarchaeological and archaeobotanical datasets. All three loors were gridded into 50cm grids and 1 L samples taken from the squares, in the case of Room A and Room B from alternate squares in a checkerboard fashion, in Room C from every single square. While it may be assumed from both size and coniguration that Room A and Room B were internal rooms, it seems likely that Room C was part of a courtyard: its loor comprised compressed grey clay containing a large amount of charcoal, a lot of ceramic potsherds, fragments of mudbrick, a considerable amount of bone and a large number of both small pebbles (up to 2cm long) and larger stones (up to 8cm long). Phase U4: Lower Assyrian Phase With the removal of the loors of the upper Assyrian levels, excavation continued down to reveal evidence of an earlier phase, distinguished by a wall (U-087) in the southeastern part of the trench underlying wall U-043 together with associated loor levels in the area of Room B. he surfaces had every appearance of being outside surfaces. here was also a potstand installed into surface U-094 consisting of a hole lined with lat pebbles 4-7cm long set in clay, with the upper rim composed of fragments of baked brick; the overall diameter of this installation was 35cm and the internal diameter 23cm. here therefore seems to be every likelihood that yet earlier phases lie below, but these layers remain unexplored. To the west of Wall U-040 was a street, irst encountered as an area of both larger cobbles in the southeastern part of the trench (U-046) and an area of smaller pebbles (U-050) in the north. It became clear that U-046 overlying U-050 overlying U-051 were all parts of the same street. Together they formed a metalling preserved up to 10 cm thick. hey were laid on a band Fig. 15. Street U-084 early phase of the Late Assyrian street showing of heavy brown clay thickly the pebbled surface and a rich deposit of zooarchaeological remains. set with pebbles (U-082) and this, in turn, overlay the stone surface of an earlier phase of the street, U-084 (Fig. 15) which it may be presumed corresponds with the earlier Assyrian phase found elsewhere. At some stage Wall U-040 was faced with an extra line of bricks (U-079) on its western side, laid directly onto the street surface, most likely as a repair or reinforcement against erosion caused by water in the street. A tiny clay juglet U-078 (ZT 40279) was found in the tough pink mudbrick collapse overlying the upper phase of the street. Anatolica XLI, 2015 23 Summary: Operation U excavations Work in Operation U succeeded in the two primary objectives of mapping the entire residence by resistivity and in excavating a sample 10m by 10m square in order to both groundtruth the results of the geophysics and to gain a more detailed picture of the stratigraphy involved. While the discovery of Late Roman remains in this area came as a surprise they are an important addition to our knowledge of the history of the site and indeed important in the context of the Roman occupation of eastern Anatolia. Operation V he architectural remains in Operation V, situated slightly northwest of the southern city gate (Operation Q, cf. Matney et al. 2011), were irst detected in a 2006 electrical resistivity survey (cf. Matney et al. 2007: 48 passim). A set of two parallel rows of rooms could be identiied, measuring approximately 30m by 8m and separated by an open space, possibly a courtyard. he eastern row seemed to comprise a series of three rooms running NE-SW, with an approximate width of 6m and a total length of at least 30m. In the 2012 season work was commenced in this area under the direction of Kristina Sauer. he main objective was to conirm the geophysical map and to identify the function of the building, which was initially thought to represent some sort of storage facility or possibly military accommodation. Work began in grid square N790E880, where the corner of a pavement of black and white pebbles (V-047) was soon uncovered very close to surface. Since the area was badly afected by irrigation and ploughing, no traces of walls could be identiied with certainty in the initial square and it was decided to follow the extension of the pebble loor to the south and by so doing uncover the whole room. hree occupation levels, two of them with subphases, were thereby identiied: (1) the main Late Assyrian Level V2 with a late phase V2b of the pebble pavement and an older Phase V2a identiied in a sounding in the SE of the room; (2) a later Level V1 marked by the presence of numerous tannurs overlaying the pebble pavement with two phases V1b and V1a which, based on the pottery, also seems to be Assyrian in date; (3) an overlaying level of uncertain date marked by a number of pits (Level V0). Level V1: the tannur level In N780E870 two tannurs, V-029 and V-041, were discovered preserved to a height of approximately 0.3m and set on top of a trodden mud surface which does not represent a carefully laid loor but rather a utilitarian surface. he younger tannur V-041, Phase V1b, measuring 0.6m in average diameter is smaller than the older one (V-029) of Phase V1a which measures 0.8m. Both tannurs are constructed of solid clay in their lower parts and large Assyrian pithos sherds in their upper parts. Both had linings made of rim fragments, also re-using Assyrian pithoi. In between the two tannurs was a large grinding stone of white limestone, trapezoidal in form, with a depression in one end and a smoothed surface. he stone was cut into the underlying pebble pavement. Once the trodden mud surface around the two tannurs was discovered, a 1 by 1m grid was laid out in order to take environmental samples from both the 24 Timothy Matney et al. supraloor and the mud surface itself. A detailed study of these samples awaits future research, but preliminary evaluations suggest that both the surface and its supraloor were strikingly clean and poor in terms of archaeobotanical and zooarchaeological remains. North of this working-space a sequence of ive postholes was discovered, forming a rough square, together with a surface made of lat pebbles, sherds, and baked brick fragments. he postholes were circular in shape, measuring up to 0.5m in diameter and laid out with pebbles, each having a larger, lat stone in middle at the bottom. hese features would appear to belong to a later encampment. A single door socket was found abutting the trench’s western edge, possibly belonging to a building or room further west. he date of these features is uncertain, but since the pottery of this level is also Assyrian, it might represent a very late Assyrian or even post-Assyrian occupation in the city. Level V2: he Late Assyrian Level he main work in Operation V was dedicated to uncovering the pebble pavement of the Late Assyrian building, and to gain insights in the function of the building itself (Fig. 16). he pavement was recovered in its entirety with the exception of an area in the northeastern corner where proximity to the surface had led to damage by ploughing; there is also damage caused by a number of pits cutting into this level. As with the overlying surfaces, a 1m by 1m grid was laid out in order to obtain environmental samples from both the supraloor and the pebble pavement itself. With regard to the walls, the south wall was the best preserved, with mud brick lines and mortar joins which could be recognized, showing at least three and a half rows of mudbricks abutting the pebble pavement. In order to clarify the width of the eastern and western walls, three slot trenches were excavated whose sections showed that the western wall, measuring up to 2m wide, was noticeably thicker than the other walls, which only are 1.5 to 1.6m across. Once the wall lines were ascertained, the collapse outside the room at its eastern and southern façade was removed in anticipation of inding an outer surface. Instead, what was actually recovered was an older wall, representing Phase V2a, protruding 0.7m from the eastern wall of the room and clearly diferent in color: the mud bricks of Phase V2b are reddish and show a great amount of white chalk inclusions, whereas the older wall is made of yellowish bricks. Unfortunately, their size could not be determined. he northern wall (V-002) of the room, separating it from the small room, or rather corridor, was the worst preserved, which is not surprising considering the proximity of the deposit to the topsoil here. Since the area was also disturbed Fig. 16. Late Assyrian pebbled pavement in Level V2. Anatolica XLI, 2015 25 by a large pit (V-019) this area was chosen as the site of a deeper sounding (V-086) in a 3m by 3m square with the aim of investigating whether there were any older loor levels or buildings levels: none were actually found, however, although this may be because the area was simply too damaged by agriculture. he pebble loor itself was built on a platform of mud bricks at least four courses high, on top of which was applied a layer of grayish mud up to 10cm thick. Into this the pebbles were set vertically side by side, producing a compact, solid and resilient loor surface which was ideal for frequent use. he pebbles come from the Tigris riverbed. Most of them are roundish and lat with smoothed surfaces, and collecting them as well as setting them into the loor must have taken a considerable amount of time and labor. Further noticeable features of the loor are distinct bands of pebbles crossing the width of the room, apparently some form of partition. his kind of loor seems to be purely utilitarian in use and contrasts, for example, with the checkerboard mosaic pavements of Operation G/R: those pavements were moreover located in courtyards and certainly had a prestige value, whereas the Operation V pavement was located inside a room in a non-residential building. his raises the question as to the function of the building. he small inds recovered included a spool-shaped token (ZT 43120), an oxhide-shaped token (ZT 43174), a token fragment (ZT 43140), an iron spearhead (ZT 43188), a fragment of a bronze ibula (ZT 43360), a large red banded agate bead (ZT 43357) and a small blue frit bead (ZT 43106). hese clues are not decisive but perhaps hint at a military and administrative function. No doorways leading to the room could be identiied since the postholes and the door socket mentioned above clearly belong to a later phase. here is a noticeable recess in the northwest part of the pebble pavement which probably indicates a doorway to the open space west of the room, but here as well no sign of an actual doorway could be traced. he other question is that of rooing. he room measures more then 5m in width and 13m in length. As the maximum span of a single beam is unlikely to be in excess of 4.5m (cf. Miglus 1999; Stone 1987, 53) this space must have had some form of central supports. As no postholes or pillar bases were discovered in the middle of the room, the supports must have been placed immediately onto the pebble pavement. Only a single feature in the northwest of the room could indicate such a feature: two massive limestone slabs set into the pavement, whose function otherwise remains obscure. Grave V-065 In the southern part of the room were discovered the poorly preserved remains of a disturbed burial. he individual was laid in a shallow pit cut into the pebble loor, lying in a crouched position on his left side, with his head at the west end facing north. he arms were crossed in front of the upper chest. he upper part of the grave had been disturbed by ploughing while the lower part of the body, from the lower chest onwards, was cut away by a large pit (V-066); in these circumstances it is not surprising that no grave goods were found. East of the main excavation area in Operation V two small soundings were excavated in grid square N780E890, the southwestern one covering an area of 3m by 3m, the northwestern one 2m by 2m. Both revealed a cobbled surface, of which the northeastern one is situated 26 Timothy Matney et al. 0.7m deeper than the other. hese stones, which must be the source of the anomalies visible in the resistivity map, presumably represent a metalled surface in the vicinity of the city gate. Summary: Operation V excavations Investigations in Operation V once more proved the accuracy of the electrical resistivity survey carried out at Ziyaret Tepe. Furthermore, they revealed part of a building which seems to be purely utilitarian in function with a pavement which must have required a considerable investment of labor to construct and whose purpose must have been related to the function of the building. Since the small inds recovered do not greatly elucidate the room’s function, the results of the scientiic analysis of the intense sampling undertaken during excavation must be awaited. Considering the layout of the whole building as visible in the electrical resistivity map, and the close proximity to the southern city gate, it seems quite probable that the building may represent a large storage complex, even though no traces of storage jars were found except for the sherds reused in Level V1. Another possible interpretation, based on the fact that the building appears to extend west beyond an open space and on to a second row of rooms, is that it was some sort of caravanserai. Fig. 17. Late Assyrian architecture in Operation G/R, 2001-2010. Anatolica XLI, 2015 27 Operation W Operation W is the designation given to some renewed investigations into speciic features in the administrative complex excavated between 2001 and 2010 as Operations G/R (Fig. 17). In 2012 two trenches were excavated, comprising a re-investigation of Courtyard 11 (Trench 1) and a re-investigation of Room 10 (Trench 2). In 2013, work was resumed in Operation W with the aim of re-investigating Room 9 (Trench 3). Trench 1 he irst trench was sited in the area of Courtyard 11. his is the southernmost of the two courtyards of Building 2, measuring 11 m (east-west) by 11.5 m (north-south). Courtyard 11 was one of three courtyards paved with black and white river pebbles laid in a checkerboard mosaic found in Operation G/R. On the whole, the pavement of Courtyard 11 was neatly laid but there were four areas where the pavement had been cut into and then relaid with much larger and rougher blocks (G-341, G-342, G-343, R-298). Our working hypothesis was that these were graves. In order to test this, we decided to excavate one of these features, choosing G-341 for this purpose. he stones were carefully removed and assembled in a cut of the same dimensions marked out on the surface so that they could be relaid correctly upon conclusion of the work. he cut into the pebble mosaic was shored up with thin wooden planks secured by iron pins in order to prevent stones from the edge of the cut falling away. his removal of the upper stones revealed a layer of hard brown clay 30cm thick, initially devoid of inclusions but in its lower part containing pebbles Fig. 18. Lower pebble pavement in Operation W, Sounding 1. and some potsherds. his then came down onto an irregular pebble/cobble surface which in turn immediately overlay a much more substantial cobble surface which was well laid though not arranged in checkerboard squares in the manner of the upper pavement (Fig. 18). his was clearly the pavement of a major earlier phase and we decided to conclude the investigation at this stage, not wishing to cut into the earlier pavement and having demonstrated, at least, that these irregular features cut into Courtyard 11 are not graves. What they are remains unclear. Perhaps, surprisingly, they are simply repairs, if not overly elegant ones. Nevertheless this investigation produced an important result, the demonstration that a substantial earlier phase underlay the Operation G/R complex. 28 Timothy Matney et al. Trench 2 he second trench excavated in Operation W was just to the east, sited in order to come down onto Room 10. his is the larger of the two rooms where tablets were found in earlier seasons. Based on the discovery from the irst trench that there was a major earlier phase to the building, the aim of Trench 2 was to investigate whether evidence for administrative use could be recovered from this earlier phase. After removing the backill, we excavated the remaining parts of the loor levels as they had been left in 2002 (north end of the trench) and 2003 (south end of the trench). he loor matrix consisted of red clay with patches of pebbles and cobbles set in, as well as fragments of bitumen (ZT 44029) and some patches of ash. In this matrix we found a cuneiform tablet (ZT 44030) and a possible sealing (ZT 44031). Removal of these surfaces brought us down onto a subloor constructed of pink clay (W-011), which had a door socket (W-009) in situ on the western side. his was a very welcome discovery as the location of the door into Room 10 is something which we had not been able to establish in earlier seasons. Removing this packing exposed another surface (W-015), which consisted of a layer of broken mudbrick overlying the loor W-016, into which a pit for a pithos (W-013) had been cut, the rim was lined by a square of baked half-bricks (W-012). he soil from inside the pithos was taken and bagged for lotation and it is hoped that this may give information on the contents of the pithos and it will be interesting to see how this compares with the information derived from the tablets found in earlier investigations. he capacity of the pithos was 180 L. All these upper layers were associated with the later level of the complex. Excavating down further we reached the wall (W-019) and upper loor level (W-018) of the expected lower building phase. Below that loor level was a level of clay packing (W-020) which in turn overlay an even lower loor level (W-021). his lower level, which is on the same elevation as the lower cobble surface found in Trench 1, had a small door socket in situ (W-022) on the western side and was associated with a large number of pithos fragments, further evidence of the protracted use of this space in this capacity. Trench 3 In 2013 work was resumed in Operation W, this time focusing on Room 9, the smaller archive room of Building 2, with the aim of recovering a more complete picture of the architectural history of the complex. he removal of the backill from the 2002 excavation (W-026) produced a spherical black bead (ZT 44090) while the cleaning of the loor as left in 2002 (W-027) yielded four baked clay tokens (ZT 44095). Below this was found a sub-loor packing comprising stones and large sherds (W-028) overlying a band (W-030) of ashy dirt mixed with clay and containing bones, sherds, brickbats and a baked clay disc (ZT 44109). his band overlay the solid foundation of the loor constructed of mudbricks of tough red clay (W-031). Removing this we came down onto the foundations of the walls of the upper phase which consisted of three courses of grey brick (W-035, W-036). hese foundations were laid, somewhat irregularly, directly on top of the levelled walls of the underlying phase (W-032, W-037, W-038), indicating that the upper architecture was a rebuilding along the same lines as the lower and with no signiicant interval of time: in other words, the upper and lower phases together form one level. Salient to this point is that the pithos G-723 was clearly installed as Anatolica XLI, 2015 29 part of the earlier phase, but kept in use in the later phase as well. he walls of the lower phase deined a space measuring 1.75 by 3.5m, considerably smaller than the 2.95 by 3.60m of Room 9 above. Where exposed, these walls were faced with half-bricks measuring 40-42cm by 20-21cm; a bench or low platform edged with mudbrick ran along the western side of the room. A mudbrick platform W-040 measuring 0.80m wide and at least 1.50m long was built against the northern wall (W-037). Surface W-039 ran underneath this platform, indicating that it was a secondary installation. Below this surface was a levelling ill (W-041), a solid packing of broken brick, yielded a clay sealing (ZT 44116) and numerous lecks of red pigment (ZT 44133). Underneath all of this was a well preserved wall (W-046) of a yet earlier phase running north-south across the trench. he material on both the east (W-045) and the west (W-049) side of the wall comprised a very solid mudbrick packing with a trampled surface (W042). No other associated surface was found, so that it appears that wall W-046 is preserved foundations only (to a height of 30 cm). here was, however, a surface of very tough grey clay with a large admixture of pebbles which ran underneath the wall, exposed on both the east (W050) and the west (W-052) sides: this must either represent the surface of an even earlier phase or, perhaps more likely, a working surface laid down as part of the initial preparation of the site. Operation Y Operation Y lies at the southwest edge of the lower town, to the west of the gateway in the city wall excavated at Operation Q and was excavated by Dr. Mary Shepperson (Matney et al 2009). It was intended to target an area of architecture, identiied through geophysical survey, which lies adjacent to a point on the city wall where the wall appears to make a ninety degree turn. he city wall makes a further ninety degree turn in the opposite direction a short distance to the south west, to form a southward jog in the otherwise fairly uniform wall. here were indications in the geophysical survey of a large public building lying just inside the wall at Operation Y, which it was thought might relate to this interruption in the course of the city wall. An initial area of 20m by 10m was opened in 2013, comprising the grid squares N790E820 and N800E820. After the removal of the topsoil this area was reduced to 15m by 10m, abandoning excavation in the northern half of square N800 E820 due to time constraints. he recovered archaeology belonged to two main phases of activity: the Late Assyrian city wall and its associated architecture (Level Y2) and a later, probably Medieval, phase of pits, cobbled features and graves (Level Y1). Level Y1 Below the plough zone, a mixed ill of redeposited building and occupation debris was cut by a series of shallow pits of various size and shape. he largest pit (Y-011/Y-022) was roughly oblong, almost 4m in length and 50cm deep. Its shape suggests it was possibly two circular pits joined together. he ill (Y-012) contained a medieval glazed sherd suggesting the pit dates to this period or later. here was no obvious order to the layout of the pits and little can be said about their function. here was no identiiable surface from which they were cut 30 Timothy Matney et al. and the material with which they were illed contained only moderate amounts of pottery and animal bone with no other inds. he only variation was in pits Y-020 and Y-026, which were found at a somewhat lower depth to the other pits and cut into the Assyrian architectural deposits below. hese pits had an ashy ill with a higher concentration of pottery and animal bone, probably originating in the ashy layer Y-034 into which both pits intruded. It seems likely that Y-020 and Y-026 may be earlier in date than the other pits, probably being dug not long after the destruction of the Late Assyrian architecture. A large cobble lens Y-002 was found just below the plough zone by the eastern edge of the excavation. he excavated portion covered a semi-circular area approximately 3.5m wide which sloped gently downwards from south to north. It was formed of irregular rounded river cobbles of various sizes and had no irm edge but rather diminished from a central concentration, suggesting a dump of cobbles rather than a constructed surface or foundation, deposited on the debris which sloped down from the raised area of the underlying city wall. he inal elements of the late phase of activity at Operation Y are four poorly preserved human burials found in the southeastern corner of the excavation, overlying and slightly cutting the remains of the Late Assyrian city wall. he four burials (Y-010, Y-013, Y-014 and Y-015) were very homogeneous in character; all were adults buried in an extended position, oriented east-west with the head at the west end and tilted onto their right sides so that they faced southward. he four graves formed a well-aligned rectangle, suggesting the graves were formally arranged in lines and rows and probably extend to the south of the excavation area. No burial goods were found in any of the four graves. Given their orientation, posture and lack of grave goods, a Medieval date seems most likely. Level Y2: Late Assyrian Phase he main feature of the Late Assyrian period architecture in Operation Y is the lower town’s city wall (Y-024). his intrudes into the southern half of square N790E820, running southwest-northeast from the southwest corner of the square, making a right-angled turn and running southeastward into the eastern section of the square. he city wall is very substantial at this point and, although the full width of the wall does not lie within the trench, it is at least 4.70m wide at this point. he manner in which the wall makes its ninety-degree turn is unusual. Rather than making a square corner by laying the square (35cm x 35cm) bricks in straight rows, the corner is constructed as a smooth curve with concentric rows of bricks fanning out to make the turn. On the inside of the curve, which is the exterior of the city wall, the bricks appear to start fanning in the opposite direction, suggesting that the angle of the wall was reinforced with curving brickwork to form an almost circular junction (Fig. 19). A remote possibility is that this could be the base for a circular tower or buttress, but this remains unlikely in the rectilinear context of Late Assyrian architecture. he bricks of the city wall vary in color, composition, and quality. he majority of the bricks are fairly well made from a ine, hard, pale greyish-brown, silty clay, but poorer-quality bricks are interspersed within the wall, made of courser reddish and greyish clays and softer, siltier material. he mortar which binds the bricks together also varies but is most commonly a ine grey mud plaster, although a material similar to the course, reddish clay found in the nearby natural deposits is also frequently found. Anatolica XLI, 2015 31 A possible explanation of why the Assyrian builders chose to construct a curved rather than a square corner is that a roadway was intended to run around the inside of the wall. An extremely hard-packed clay-rich deposit (Y-031) runs against the curve of the wall in a 2.40m wide strip. he heavily compacted clay contains frequent small-medium sized pebbles and a moderate number of potsherds, making it a very hard, durable construction. he placement of a roadway along the inside of the wall would have brought traic to and from the Operation Q gate and in emergencies it would have allowed troops and supplies to move along the city wall for the better defense of the gateway. A curved bend rather than a sharp corner here would have greatly eased the pasFig. 19. Late Assyrian architecture, Level Y2. sage of people, animals and wheeled vehicles. It is interesting to note that a similar gap between the city wall and the adjacent architecture was present in the area excavated in Operation K. he only other element of Neo-Assyrian architecture preserved in Operation Y is the mudbrick wall (Y-029) of a building, which runs parallel to the northwest internal face of the city wall in the western part of the trench. At about the point where the city wall turns back to the southeast, wall Y-029 makes a right-angled corner turning in the opposite direction towards the northwest, although only traces of this corner survive. he mudbricks at the western end of wall Y-029 are very well preserved and show the wall to be made of two rows of large complete mudbricks, 40cm x 40cm square, laid in a ine plaster mortar, making a wall 85cm wide. As they run eastward the bricks become less well preserved, becoming only traces of brick material by the point where the wall turns the corner, after which the line of the wall is only preserved as a foundation cut. It seems that only one course of this wall survives, suggesting 32 Timothy Matney et al. that the building was deliberately levelled after it fell out of use. he foundation trench for wall Y-029 is cut through a deposit of tough reddish clay with frequent white chalky inclusions. his deposit, which is found on both sides of wall Y-029, seems to be re-deposited natural clay used to prepare a level area for the construction of the Y-029 building. his levelling layer was probably necessary because the deposit below, which runs under wall Y-029, was a soft ashy ill (Y-034) mixed with mudbrick lumps, rich in pottery and animal bones and generally unsuitable for the foundation of a building. It seems that after the completion of the city wall and the roadway (Y-031), the area immediately inside the wall consisted of mixed building rubble and ashy rubbish dumps. his area was then levelled and the ashy deposit capped with the layer of reddish clay (Y-030) before the foundation of wall Y-029 and the rest of this building was cut into it. In a later phase, this building seems to have been destroyed and levelled leaving only these traces of its foundation. here is no trace of later building activity so the area was probably left as open space inside the city wall. Summary: Operation Y excavations Although little of the building inside the city wall remained, the organization of the Assyrian architecture here displays some important features. Most notable is the emphasis on defensive features at this key point near a major gateway. Firstly, the double bend in the city wall, of which the wall at Operation Y forms one half, seems to serve little purpose except to provide a stretch of wall between the two corners which faces towards the Operation Q gateway; this would have allowed defenders a better angle to aim arrows, sling stones, and any other projectiles down onto the lank of an enemy attacking the gateway. he provision of a reasonably wide roadway along the inside face of the wall also seems to be aimed more towards the movement of troops and equipment along the city’s defensive perimeter than the general movement of goods and people, which in most contemporary city plans tends to be primarily radial, between the periphery and the centre. he care taken to construct a curved corner in the city wall at Operation Y suggests that the easy movement of traic along this route was considered of some importance in order to merit such an unusual arrangement. he width of the wall and the extra buttressing at this corner also suggest a signiicant concern with defense here. Defensive concerns may even possibly account for the destruction and levelling of the building inside the curve of the wall, which is unusual in the lower town where demolished buildings are generally built over with new structures. Even with the unusual curve, the right-angled turn in the roadway behind the city wall would have constituted an awkward bottleneck where traic was forced to slow to take the turn. It is possible that the building inside the wall was not rebuilt in order to provide greater space here for people, animals and vehicles to pass and make movement along the wall easier. Operation Z Operation Z, the last of the archaeological investigations to be undertaken at Ziyaret Tepe, was the designation for a 5m by 5 m trench located in grid square N810E840, a short distance northeast of Operation Y. he aim of Operation Z was to ground truth the existence Anatolica XLI, 2015 33 of a building whose outline appeared on the resistivity plot of the area from 2006 (Matney et al. 2007: 48, 73 ig. 22). Excavation was carried out in the 2013 season directed by Dr. John MacGinnis and assisted by Daniella Arroyo. Having removed the topsoil, the irst feature encountered in Operation Z was a wall running northeast-southwest across the southern half of the trench, preserved for a length of 3.5 m, with a curious facing of stones and potsherds set on edge (Z-002), and with an associated loor with a thin red plaster surface (Z-008) laid on a sub-loor packing (Z-010) of red clay on the southeastern side. he stone and pottery facing lay directly against a larger mud brick (wall Z-005) which belongs to an earlier phase of occupation. his wall was 1.80m wide and made of bricks measuring 38-40cm square. he wall was not very well preserved and in some places diicult to follow, but in any case deined on the southeastern side by the red plastered surface Z-008 and to the northwest by a rough stone pavement containing fragments of plaster, pottery and animal bones (Z-009). Underneath this was a zone of alternating bands of broken red brick and layers of ash. Below this level, a stone foundation (Z-014) was identiied running NW-SE and continuing under wall Z-005, and to the north of this were the fragmentary remains of a white plastered loor (Z-019). he work in Operation Z succeeded in its primary aim of ground-truthing the architecture visible in the resistivity mapping of 2006, which appeared as a substantial wall exactly where expected. his wall was clearly Late Assyrian in date. In addition to this, the work in Operation Z yielded evidence for a subsequent occupation of a more modest nature which reused the main Assyrian wall and must as a consequence be either Late Assyrian or immediately post-Assyrian in date. Archaeobotanical Reports As noted above, systematic collection of paleofaunal and zooarchaeological samples was standard procedure at the Ziyaret Tepe and an exceptionally rich and robust dataset exists for most Operations and time periods. Below, the preliminary results of two studies are reported: a summary of the Late Antique samples from Operations T and U by Lucas Proctor and the Late Assyrian samples from Operation Q, excavated prior to 2011 by Dr. Melissa Rosenzweig. Sample processing and identiication followed methods previously established for the Ziyaret Tepe archaeobotanical assemblage by Rosenzweig (Matney et al. 2011). All sediment samples were processed using a Siraf-type lotation machine by Rosenzweig and Mr. Suphi Kaya, using 1mm and 355μm mesh to recover heavy and light fractions, respectively. In order to facilitate easier sorting and identiication, all samples were divided into four size grades using 2mm, 1mm, and 500μm nested sieves and larger samples were further subdivided using a rile splitter. Grades greater than 1mm were fully sorted, while grades less than 1mm were scanned until no further identiiable specimens could be retrieved. Fragments of wood charcoal were separated and weighed in grades above 1mm and saved for later analysis. A Leica EZ4 optical microscope with a maximum magniication of 35× was used for all sorting and identiication. Plant remains were identiied to the genus or species level (when possible) based on Rosenzweig’s modern reference collection and the UConn Archaeobotanical Reference Collection, as 34 Timothy Matney et al. well as several reference guides, including Nesbitt (2006), Cappers, Neef, and Bekker (2009), Martin and Barkley (1961). After identiication and quantiication of the sorted material from all samples, analysis of the assemblage was conducted using simple quantitative measures, including ratios, density, diversity, proportion, and ubiquity, commonly applied in archaeobotanical research (Miller 1988; Popper 1988; Pearsall 2000). Late Antique Archaeobotanical Report (L. Proctor) his report presents the results of an archaeobotanical analysis of twenty-nine samples from the 5th to 8th centuries AD at Ziyaret Tepe. While comparatively little research has been conducted on archaeobotanical remains from Late Antique contexts throughout Anatolia (Marston and Miller 2014), historical and textual evidence has highlighted a dramatic population growth supported by heavy economic development and intensive agricultural production during late antiquity, especially in Southeastern Turkey and the Levant (Decker 2007; 2009). Survey work conducted by Algaze in the upper Tigris River Valley has supported this by inding evidence for heavy settlement along the Tigris River during this period, with sites lining both Fig. 20. Description of all samples examined, including pre-lotation soil volume (L), Total number of specimens (NSP), density of remains (NSP/L), and Shannon-Weaver (SW) diversity index (see Popper 1988). * denotes samples that are sterile or contain only indeterminate remains. Anatolica XLI, 2015 35 banks from Batman Su to the modern city of Bismil. (Algaze et al. 1991: 184). he following analysis situates the Late Antique occupation at Ziyaret Tepe within the context of regional agricultural and economic intensiication and population growth during this period, despite its comparatively small size. Twenty-nine Late Antique sediment samples were collected for archaeobotanical analysis during the 2011-2013 excavation seasons at Ziyaret Tepe (Fig. 20). Sediment sampling strategies followed the protocols established by Matney and Rainville in 2000. Samples were collected from pit and loor contexts in volumes ranging between 0.05 and 33 L. Twelve sediment samples were collected by Dr. Kemalettin Köroğlu during excavation of Operation T. he Late Antique component of Operation T consisted of a building (T-031/T-064) with an adjacent cobble-surfaced courtyard (T-047). Samples were taken from associated surfaces, as well as three large pits. Similarly, sixteen sediment samples were collected by Dr. John MacGinnis during the excavation of Operation U. Operation U contained two phases of construction with a northern building (U-004, U-020, U-037) separated from exterior surfaces by two walls. One additional Late Antique sample was recovered from Operation K in 2013, which contained the contents of a ceramic drain pipe. Sample Composition he Late Antique samples yielded 2368 specimens, of which 1126 were identiiable plant remains (Fig. 21). he remaining 1242 specimens were heavily damaged, indeterminate fragments of carbonized remains which could not conidently be assigned to any taxonomic group. Poor preservation limited the identiication of specimens in all cases, and, overall, intact and well-preserved specimens were rare. Sample density was generally low, with a median of 8.71 NSP/L (Fig. 20). Most samples contained only a small number of identiiable taxa, though two samples, ZT 39293 and ZT 40269 both contained over 200 identiiable specimens. Sample diversity, an index of richness of and evenness of species within samples (Pearsall 1983; Popper 1988), is generally low, indicating that taxa occur both infrequently and in varying quantities within samples (Fig. 20). he overall proportions of major taxa categories (cereal, weeds, etc.) are broadly comparable between Operations T and U despite being separated by nearly 300 meters and containing diferent numbers of pit and surface contexts. Pit and loor contexts had unique archaeobotanical signatures, indicative of the diferent depositional histories and formation processes unique to these context types. Pit deposits are typically the result of continual amalgamation of multiple activities, such as crop processing, hearth clearing, and trash burning, while surfaces ought to contain limited quantities of remains due to both limited depositional events and the deleterious efects of cleaning and trampling (Miksicek 1987; Hubbard and Clapham 1992; van der Veen 2007). hirteen of the samples represented loor contexts. here were only three surface samples collected from Operation U, while there were ten from Operation T. Samples within the structure in Operation T (T-031, T-064) contained a greater proportion of weeds and non-cereal crops with no cereal chaf, while surface samples from the courtyard (T-046) contained cereal chaf and grain but few economic plants. It is notable that a clear diference in composition is present between the interior of the structure, which contained non-cereal crops and a greater proportion of weedy taxa, compared to the samples from the courtyard 36 Timothy Matney et al. which contained crop processing debris and a greater proportion of cereals (Fig. 22a). Pits represented the majority of the samples collected and were comparatively the richest contexts in assemblage. Samples from pit contexts yielded, on average, 144.4 specimens per sample, though sample richness was uneven (Diversity: Low = 0.80; High = 2.08). Compared to loors, pits contained a greater proportion of cereal remains and lower proportions of pulses, other economic taxa, and weeds. One sample from Operation U, ZT 40269, was exceptionally rich in botanical remains. ZT 40269 was taken from pit U-071, located between the two Late Antique walls. It contained 687 specimens, 317 of which were identiiable, which makes up nearly 30% of the entire assemblage. he identiiable specimens were composed of 75% cereal grains, as well as 23% weeds and 2% of chaf. Weeds were particularly diverse, containing 24 taxa. A small burnt bead was recovered from the sample, which could suggest that the contents of the pit were burned in situ with other trash. here are many possible routes leading to the inclusion of plant remains in the archaeological record, including food use/discard, crop processing activities, and the use of dung as fuel (Miller and Smart 1984; Charles 1998; van der Veen 2007). It is probable that some combination of several of these processes contributed to the formation of this assemblage. However, it is argued here that this assemblage primarily derived from the burning of dung as a supplement to scarce wood fuel resources (following Miller and Smart 1984; Miller 1996; Charles 1998), as evidenced by a high seed/charcoal ratio (97.27 ct/g) and low wood charcoal densities (average of 0.16 g/mL of material). While there were few dung fragments recovered from the specimens, the use of dung fuel is ubiquitous in the area today and there is evidence that it was used in earlier Assyrian contexts at Ziyaret Tepe (Rosenzweig 2014). Furthermore, the heavily damaged nature of the recovered plant remains may be indicative of partial digestion. However, the overall cultigen/non-cultigen ratio, which has been used as an indicator of dung fuel usage (Miller & Smart 1984) is very high (3.70), contrary to expected results for dung fuel usage. Results Cereal grains were the most abundant plant remains, making up 57% of the identiiable assemblage (Fig. 22b). Wheat (Triticum sp.), including both free-threshing (T. aestivum/T. durum), and some hulled (T. monococcum/T. dicoccum) varieties, were most abundant (n = 58) and occurred in 36% of samples. Free-threshing wheat greatly outnumbered hulled wheat, suggesting that it was the preferred wheat crop (Table 2). Barley (Hordeum sp.) has 28% ubiquity and is as abundant as free-threshing wheat. In addition to wheat and barley, oat (Avena secale.) and rye (Secale sp.) both appear in the assemblage, albeit in very small quantities; only three oats and one rye grain were recovered. Large, indeterminate cereal grain fragments constituted the bulk of the cereal specimens (n=548) and were found in 80% of samples. Given that free-threshing wheat and barley were the most commonly encountered cereal varieties, it is likely that most of the indeterminate fragments come from these taxa. Cereal processing debris and chaf contributed to 17% of the identiiable assemblage (n=193). Wheat chaf was the most abundant and frequently encountered chaf, having a ubiquity of 36% of all samples, and included Triticum sp. rachis, Triticum sp. spikelet fork, Triticum sp. glume/base, T. mono- Anatolica XLI, 2015 37 coccum spikelet fork, T. durum rachis, T. aestivum rachis, and indeterminate free-threshing wheat rachis. Triticum sp. rachis (n=26) and free-threshing wheat rachis (n=22) were most abundant types of chaf. Hordeum sp. processing debris was also present, but occurred in only three samples. Free-threshing wheat and barley appear to have been the primary staple crops produced at Ziyaret Tepe during the Late Antique. Free-threshing wheat would have been produced exclusively for human consumption, while barley may have been produced for human consumption as food or beer, fodder for animals, or a combination of both (Miller 1996). he degree to which hulled wheat, rye, and oat were locally consumed/produced at Ziyaret Tepe is unknown. heir limited presence in this assemblage is curious, but little can be said about their importance without further evidence. Economic legumes constituted a very small percentage of the identiiable Late Antique plant assemblage at Ziyaret Tepe (Fig. 22b). A total of 22 specimens of pulse crops (2% of the identiiable assemblage) were found in nine samples (Fig. 21). Most of the Late Antique specimens came from pit features, likely as part of burnt refuse or hearth sweepings. Bitter vetch (Vicia ervilia) was the most abundant pulse, with twelve specimens overall, eleven of which were found in ZT 39293. Bitter vetch is a common and early domesticate in southwest Asia, though its use as an economic taxa has sharply declined. Today it is primarily used as animal fodder, but in the past, bitter vetch was used as a famine crop, because of its bitter taste and toxicity (Helbaek 1961; Zohary et al. 2012). Other legumes recovered include a single specimen of pea (Pisum sativum) from a pit in Operation U and a chickpea (Cicer arietinum) from the ceramic pipe in Operation K. Six indeterminate large legumes were also recovered. Economic taxa other than cereals and pulses contributed to less than 2% to the identiiable plant assemblage. Of these taxa, indeterminate nut shell fragments were most abundant (Fig. 21), however, the shell was too fragmentary for any speciic taxonomic identiication. hree specimens of grape (Vitis vinifera) were recovered from three samples. Each specimen occurred individually, and in diferent contexts (loor, pit, and drain pipe). Fig (Ficus sp.) occurs very infrequently in the assemblage, being represented by only three specimens from two samples. Finally, one mineralized hackberry (Celtis sp.) was also recovered from a ceramic drain pipe found in Operation K. Weeds taxa contributed, in total, 22% of the identiied plant assemblage, and were present in all but two of the analyzed samples. here were 47 identiied taxa from a variety of diferent families, genera, and species (Fig. 21). Individual taxa occurred infrequently and in low numbers. Much of the weed data derived from ZT 40269, which produced 24 diferent weed taxa. hroughout the Late Antique assemblage, the most commonly recovered taxa were steppe weeds, which comprised 52% of the weed assemblage and 11% of the total identiiable assemblage. Timothy grass (Phleum excaratum) was the most frequently occurring steppe weed (48% ubiquity). One unknown grass species, “Grass 001” was also frequently recovered (40% ubiquity), often co-occurring with timothy grass. Steppe weeds grow in the dry slopes and uplands surrounding the Tigris River Valley, rather than near to the settlement or in ields. hus, the most likely vector for steppe weeds to have been deposited in the archaeobotanical record is through the pasturing of livestock in these areas, which further supports the use of dung fuel. 38 Timothy Matney et al. Fig. 21. Summary of identiiable taxa from Late Antique samples. Number of identiiable specimens (NISP) and ubiquity (percent of samples containing taxa) are reported. Anatolica XLI, 2015 39 Wetland adapted weeds, such as Cyperus sp., Eleocharis sp., and Carex sp. were recovered in 28% of samples, but comprised only 8% of the total weed assemblage (Fig. 21). Wetland adapted weedy taxa often have been used as evidence for irrigation (Marston 2011; Marston and Miller 2014: 766-767). However, while wetland taxa were present in the Late Antique assemblage, they occurred in insuicient numbers to state conclusively that irrigation was practiced. Irrigation canals have yet to be observed in the vicinity of Ziyaret Tepe, though irrigation is ubiquitous in the Tigris Valley today and Late Antique irrigation works have been observed regionally (Decker 2007). Field weeds comprised 10% of the total weed encountered. Bedstraw (Galium sp.), a large and easily recognizable ield weed commonly encountered in archaeobotanical assemblages, was observed in three of the samples. hree Cephalaria sp., a bitter, noxious weed were recovered from ZT 40269. Miller (1998: 240) and van Zeist and Bakker-Heeres (1985:254) note that Cephalaria is not commonly encountered in archaeological deposits before historic periods, becoming increasingly common in Classical and Medieval contexts. Large ield weeds, such as Cephalaria and bedstraw would have been handpicked from grain stores, because their large size and shape prevent their removal during bulk crop processing activities (cf. Jones 1984). hey could have either been directly discarded or given as fodder with other crop processing remains. Finally, 30% of the weed assemblage can be categorized as catholic/non-diagnostic taxa (Fig. 21). hese taxa are found across ecological zones, and many can be considered ruderal. Specimens belonging to the Cruciferae (mustard) and Lamiaceae (mint) families are well-represented in the assemblage. Both families were present in 20% of samples. Many members of these plant families have culinary and medicinal uses. However the low abundances limit any discussion of possible uses as herbs (cf. Behre 2007). Summary: Late Antique Archaeobotany he evidence presented in this study indicates that wheat was the major staple crop produced at Ziyaret Tepe during the Late Antique occupation. he ratio of wheat to barley (1.61) contrasts sharply with the results from the Late Assyrian (ca. 900-600 BCE) occupation at Ziyaret Tepe, where barley was strongly favored (Rosenzweig, 2014). his suggests that an important shift in local land use patterns occurred during the Late Antique period. According to contemporary Byzantine written accounts, free-threshing wheats, such bread (T. aestivum) and macaroni wheat (T. durum) were the most desirable and economically valuable variety of grain in the Late Antique East, followed by hulled wheat (see Decker 2009: 80-112). Barley, which was the second most frequently encountered economic taxa in the Late Antique assemblage, was considered inferior to wheat by Byzantine authors because it produced bread of poor quality. (Decker 2009: 80-112). hus, with the preponderance of wheat in the Late Antique samples examined, it seems unlikely that barley would have been intended for human consumption. Barley may have been produced as a fallback crop, however, given its lower water requirements, in order to mitigate the risks associated with droughts and crop failure. Ethnographic accounts from the Eastern Mediterranean have demonstrated that free-threshing wheat and barley were often grown together as a maslin crop, in order to protect against total crop failure (Jones, 1990; Jones and Halstead 1995). However, this interpretation is diicult 40 Timothy Matney et al. to prove (e.g. van der Veen 1995), and the data presented here are inconclusive. A more convincing case can be made for the use of barley as animal fodder. Low ratios of weeds to cereals has been previously argued as evidence for livestock foddering, assuming dung fuel usage is also evident (Miller 1996; Marston 2011). At Ziyaret Tepe, the overall ratio of weeds to cereals is quite low (0.280), which supports the foddering of animals, since many of the remains are assumed to be coming from dung fuel sources. Supplementing grazing with barley and chaf would allow for the production of livestock above the carrying capacity of the available land. his would have a useful strategy for dealing with high regional population levels and pressure on land availability, as has been suggested by Decker (2009: 29-79). A. B. Fig. 22. a: 100% stacked bar chart depicting proportion (%) of major taxa categories by contexts. n = number of specimens; s = number of samples. b: Pie chart summarizing the major categories of taxa identiied in the Late Antique archaeobotanical assemblage. Anatolica XLI, 2015 41 Non-cereal crops, including sweet pea (Pisum sativum), bitter vetch (Vicia ervilia), chickpea (Cicer arietinum), hackberry (Celtis sp.), ig (Ficus sp.), and grape (Vitis vinifera), may well have been important parts of local diet. Because the depositional and taphonomic processes that create charred botanical assemblages frequently select against the preservation of these taxa (van der Veen 2007), the relative importance of non-cereal economic taxa are often underestimated. Additionally, other crops, such as olive, lax, melons, and dates, are conspicuously absent, despite their established economic importance in the Eastern Mediterranean both before and during Late Antiquity (see Zohary et al. 2012). hese crops are ubiquitous in the Tigris Valley today, and are represented in many of the other Late Roman/Byzantine sites throughout the Eastern Mediterranean for which there are published archaeobotanical data (Helbaek, 1961; Ramsay, 2010; Kaptijn et al., 2013; Ramsay & Smith, 2013). he Ziyaret Tepe assemblage, as a whole, displays considerably less variety in terms of economic taxa than other sites from the same time period. his lacuna may hint at the socio-economic realities of the inhabitants of the small settlement at Ziyaret Tepe, who, in order to navigate high taxes and demand for land created by a dramatic expansion of population in the Eastern Mediterranean (Decker 2009: 29-79), may have chosen to focus their productive eforts on intensively cultivating high-value wheat crops at the expense of crop variety. In conclusion, these results provide important insights into the plant use and crop husbandry of Late Antique settlements in Southeastern Turkey. Emphasis on the production of highly valued free-threshing wheat and the foddering of livestock during late antiquity suggests a dramatic shift in land use at Ziyaret Tepe to a high-risk, high-reward strategy concurrent with a regional lorescence of economic activity during late antiquity, as argued by Decker (2009). Foddering and indirect evidence for irrigation suggest a variety of possible means by which inhabitants could have attempted to increase the productive output of their land. his intensiication can be interpreted to be the result of the economic pressures placed on small local communities by a large regional population in the Eastern Mediterranean during late antiquity. Operation Q Archaeobotanical report (M. Rosenzweig) he preliminary results presented below comprise data from soil samples collected from a city gate located in the lower town of Ziyaret Tepe, on the site’s southwestern side, dubbed the ‘Khabur Gate’ by excavators. Excavation of this structure ran from 2007 to 2009 under the direction of Dr. John MacGinnis. During these three ield seasons, excavators exposed a substantial city gate structure within a horizontal area of 225m², in an area designated Operation Q. As a result of these excavations, archaeologists conidently identiied and recorded the city gate’s two-chamber design, as well as the gate’s articulation with a wide street leading into Tušhan (Matney et al. 2009: 61). Artifact associations and abutting domestic architecture link the city gate to the Late Assyrian occupation of the tell (Matney et al. 2009: 62), while it is presumed that the latest phase of the city gate culminates with the abandonment of Tušhan around 611 BCE (Grayson 1975: 95, 33-35). he ‘Khabur Gate’ served as a major corridor between the provincial center and the lands lying south of the city, which included farmsteads situated along the loodplain of the Tigris and pasturelands in the foothills beyond. hese ields and pastures were the primary loci Timothy Matney et al. 42 Fig. 23. Plan of Operation Q (Khabur Gate) showing the four phases of use. (from Matney et al. 2011: Fig. 14). of agricultural production for supplies to Tušhan’s temple and treasury, located just north of the gate, where administrators stored and redistributed the taxes in kind collected by Late Assyrian oicials (Parpola 2008; Schachner 2003). hus, the ‘Khabur Gate’ would have been a site of regular economic activity, especially the traic of agrarian goods both in and out of the settlement, and an important hub of daily life for Tušhan’s inhabitants, oficials and non-oicials alike. In support of this view, the stratigraphy of both the gate’s buttresses and transecting street reveal four phases of construction, demonstrating consistent use and renovation of this city entrance and thoroughfare throughout its use-life (Fig. 23; Matney et al. 2009: 61). Methods During excavation, archaeologists regularly sampled primary contexts of in situ deposits in Operation Q – loors and surfaces (including street), hearths, and well-deined pits – from all four construction phases, both in and around the gate. his analysis concentrates on 31 samples from the inner chambers of the gate itself, covering approximately 36m² that are composed of loor and hearth deposits (Fig. 24). he individual samples analyzed from the gate range in volume from 0.75 to 6.0 L, with a mode sample size of 2.0 L. Six samples come from the earliest phase of the gate and were collected from a burnt plaster loor in the eastern chamber. hree samples were taken from Phase II, a mud plaster loor in the eastern chamber. Unfortunately, these samples were practically sterile, producing only ive identiiable seeds. In Phase III, four samples come from hearths situated in the eastern chamber, and the other six samples are from loor contexts in the same location. All twelve samples from Phase IV come from a clay loor in the western chamber of the gate which excavators gridded into 50 cm by 50cm squares, removing two to three liters from every other square for lotation. Paleobotanical collection protocols were discussed above. Results All together, 838 identiiable plant remains were recovered from the city gate samples, representing 47 diferent plant taxa (Fig. 25). he pie chart breaks down the percentage of each plant group represented within this assemblage, with cereal grains making up 32% of the overall crop composition, followed by chaf (7%), pulses (1%) and fruit (< 1%). Weeds, deined here as non-cultigens, predominate the assemblage, comprising 60% of all the macrobotanical material identiied. Furthermore, in all likelihood most of the ‘indeterminate seeds’ represent weedy species as well, meaning that the percentage of weeds in the assemblage could reach as high as 83%. Anatolica XLI, 2015 43 his overall plant distribution relects both general human practices and taphonomic processes. Similar densities of wood charcoal and plant remains across loor and hearth samples (Fig. 26) suggest that the occupants of the gate chambers regularly cleaned out the ire installations. If the hearths had contained in situ deposits of burned material, wood charcoal and sample densities would spike in these contexts, relative to the loor samples. Instead, the low numbers of charred plant material in the hearths appear to index routine sweepings of the features’ contents. Presumably, guards within the city gate would have consumed prepared plant foods, like bread and porridge, which would not have undergone burning and ended up in the archaeological record. But when faced with relatively clean hearth debris, this hypothesis is diicult to conirm. More assuredly, the enclosed gate chambers would not have been practical locations for crop processing, and low percentages of chaf and ield weeds, even in the hearth samples (12.5% and 6.7% respectively), support this claim (cf. Hillman 1981). None of the city gate phases contained any storage vessels (Matney et al. 2009: 61-62), and lacking large deposits of clean grain, the city gate does not appear to be a site of cereal storage. Instead, it seems most likely that the assemblage largely comprises the debris of dung fuel burned in the chambers for heat, light and food preparation, even though archaeologists only found hearths ZT no(s). 2764527650 2745027452 Phase Chamber I East II East 27507 III East 27506 III East 27487 III East 27418 III East 27422 III East 27416 III East 27394 III East 27391 III East 27381 III East 27378 III East 2706827079 TOTALS IV West Context burnt plaster loor Q-261 , randomly selected points “bright mud plaster” loor Q-197, southwest portion hearth Q-166 atop mudbrick platlorm Q-164 hearth Q-165 atop mudbrick platform Q-164 sample from hearth Q-154, associated with loor Q-159 sample from hearth Q-154, associated with loor Q-155 mud loor Q-155, around pit Q-052 mud loor Q-150, northern half of chamber pebble loor Q-145, southern half of chamber plaster loor Q-144, northern half of chamber plaster loor Q-143, southern half of chamber plaster loor Q-142, northern half of chamber red clay loor Q-023, selections from 50x50 cm grid No. of samples Total sample volume (l) 6 10.1 3 3 1 2.5 1 1.5 1 3 1 6 1 3 1 3 1 0.75 1 1 1 3 1 3 12 29 31 68.85 Fig. 24. Contextual information for the Operation Q archaeobotanical samples. 44 Timothy Matney et al. Taxon Common Name Phase I Phase II Phase III Phase IV Total Count Cereals Triticum monococcum einkorn 1 1 Triticum sp. wheat 2 2 Triticum sp. spikelet fork 1 1 2 Triticum sp. rachis 1 2 3 Triticum sp. glume base 7 1 7 15 Hordeum sp. barley 26 3 15 44 Hordeum sp. rachis 6 10 8 24 Cereal grain indet. 158 22 41 221 Cereal glume indet. 1 1 2 Cereal rachis indet. 1 1 Cereal culm 5 7 12 Pulses Lens culinaris lentil 6 1 7 Pisum sativum ield pea 1 1 Large legume, indet. 1 1 Fruits Ficus carica ig 1 1 Weedy Taxa Hordeum sp. (wild) wild barley 3 3 Aegilops sp. goat grass 1 1 Aegilops sp. glume/spikelet fork 1 1 2 Eragrostis sp. love grass 1 1 2 Phleum exaratum timothy grass 20 1 104 74 199 Stipa sp. needle grass 1 3 3 7 Small grass seed indet. 45 1 10 20 76 Grass #001 3 1 1 8 13 Asteraceae daisy family 1 2 3 Anthemis sp. may weed, chamomile 1 4 5 Artemisia sp. sagebrush, wormwood 1 1 Centaurea sp. knap weed, star thistle 1 1 Apiaceae carrot, parsley family 1 1 2 Buglossoides sp. corn gromwell 16 1 17 Carophyllaceae pink family 1 1 Heliotropium sp. heliotrope 2 2 Silene sp. campion, catchly 9 9 Vaccaria sp. cow cockle 3 2 5 Chenopodium sp. goosefoot, lamb’s quarters 2 25 2 29 Hypericum sp. St. John’s wort 1 1 Cruciferae mustard family 1 1 Brassica sp. mustard 2 2 Descurainia sp. lix weed 1 1 Cyperus sp. nut grass, nutsedge 1 2 3 Astragalus sp. milk vetch 3 3 Coronilla sp. crown vetch 1 1 2 Trigonella sp. fenugreek 1 1 Trigonella astroites 7 1 8 Small legume indet. 3 4 12 19 Scleranthus uncinatus knawel 1 1 Lamiaceae mint family 1 1 Acinos sp. calamint 1 1 Mentha sp. mint 1 1 Teucrium sp. germander 1 1 Teucrium parvilora 1 1 Malva parvilora mallow 1 1 Fumaria sp. fumitory 1 1 Papaver sp. poppy 1 1 Polygonum sp. knotweed 10 4 2 16 Portulaca sp. purslane 31 1 32 Reseda sp. mignonette 1 1 Asperula sp. woodruf 1 1 Galium sp. bedstraw 3 3 Asperula/Galium spp. 4 9 3 16 Verbascum sp. mullein 1 2 3 hymelaea sp. thymelaea 1 1 IDENTIFIED REMAINS TOTALS 372 5 228 233 838 Indeterminate seed 417 30 359 424 1230 SORTED REMAINS TOTALS 789 35 587 657 2068 Fig. 25. Absolute counts of archaeobotanical remains from Operation Q by phase. Anatolica XLI, 2015 45 (n = 3) in Phase III of the city gate’s history. Nonetheless, across all phases of the city gate, the samples display low charcoal densities and low cultigen to non-cultigen (i.e. weedy taxa) ratios, both indicators of dung fuel cereals 32% burning practices (Miller and Smart 1984). n = 268 Densities of wood charcoal, which signal epweeds 60% isodes of wood ire burning, have a low avern = 501 age of 0.067 g/L, with no sample exceeding chaff 7% 0.133 g/L, even amongst the hearth samples n = 59 where wood burning would be most evident. pulses 1% Likewise, the ratios of cultigens to non-culn=9 tigens in the assemblage (0.161 average) fruits <1% n=1 compare favorably with seed recovery ratios in dung fuel cake experiments (0.014-0.031) Fig. 25 (continued). Plant group representation (ibid.), and falls signiicantly lower than the across the total assemblage of Operation Q samples. ratio (0.5) obtained from a medieval period sample at Ziyaret Tepe where wood charcoal appeared to be a primary fuel source (Rosenzweig in Matney et al. 2011). Finally, with the exception of primary hearth or storage deposits, most archaeobotanical remains derive from secondary or tertiary sources (Hillman 1981; G. Jones 1987; M. Jones 1985), and the spread of weedy taxa across time and space in the city gate chambers probably results from the regular sweeping of dung fuel waste from the indoor hearths onto the surrounding loor surfaces. To summarize, certainly there is the likelihood that the plant remains from the gate chambers represent an amalgamation of human practices. However, lacking convincing distributional trends in expectation of indoor crop processing, plant storage, or wood ire burning, the bulk of the archaeobotanical remains can be attributed to dung fuel burning, where analytical evidence does lend support. Consequently, the plant remains from the ‘Khabur Gate’ should be thought of as a composite collection of forage and fodder resources for livestock. As such, this archaeobotanical assemblage presents us with Floors Hearths All Samples information on the envi(n = 27) (n = 4) (n = 31) ronmental mosaic of ZiCrop Average Count 9.56 4.75 8.65 yaret Tepe’s surroundings, Crop Relative Abundance 34.58% 18.27% 32% as well as important clues Chaf Average Count 1.01 3.25 1.84 to land-use and manageChaf Relative Abundance 3.75% 12.5% 7% ment practices during the Wood Charcoal Density Late Assyrian period, espe0.069 0.039 0.067 (g of charcoal/L of soil) cially in regards to the proSample Density visioning of animals. 32.53 33.27 29.35 (no. of items/L of soil) From the 268 cereal grains recovered from the city gate, only 17% Cultigen to Non-Cultigen Ratio 0.130 0.085 Fig. 26. Statistical comparisons of the loor and hearth samples from Operation Q. 0.161 Timothy Matney et al. 46 A. Grain Abundance in the Total Assemblage barley 16% n = 44 unidentified cereal 83% n = 221 wheat 1% n=3 B. Grain Abundance Across Phases 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Phase I (n = 6) barley Phase III (n = 10) wheat Phase IV (n = 12) unidentified cereal Fig. 27. a: Pie chart of cereal grain representation across the total assemblage of Operation Q samples. b: Bar chart of cereal grain representation across the phases of use. here were no cereal grains recovered from Phase II (n = 3), which is not represented in the graph. could be identiied (Fig. 27a). Among those 47 caryopses, barley outweighs wheat 14 to 1. A salt-tolerant and drought-tolerant crop, barley (Hordeum sp.) would have been a reliable human staple as well as the preferred cultivated fodder for provisioning livestock in and around the semi-arid region. Meanwhile, the trace amounts of einkorn (Triticum monococcum) and unspeciied wheat (Triticum sp.), which have greater water requirements and therefore most likely relied on irrigation, would have been reserved for human consumption alone. Given that these samples are understood to primarily derive from dung fuel burning, the predominance of barley over wheat is expected, as a relection of foddering practices. Although there is a slight elevation in barley grains preserved in Phase IV of the ‘Khabur Gate’ (Fig. 27b), the Environmental Signatures Represented Weedy Taxa Environmental Signatures by Represented by Weedy Taxatrend in barley over wheat remains steadfast throughout woods <1% n=1 non-diagnostic 20% n = 97 grassland 41% n = 207 wet lands 11% n = 54 fields 13% n = 66 steppe 15% n = 76 Fig. 28. Pie chart of the abundance of weedy taxa, linked to the environmental conditions under which they are found and lourish, across the total assemblage of Operation Q samples. the use life of the gate. he continuity of grain distributions across the phases may be further interpreted as evidence of the homogeneity of cultivated foddering practices across time at Tušhan. In other words, barley appears to be the main constituent of livestock cereal provisioning regardless of phase. his claim will require further investigation into the composition of dung fuel debris across the site in the Late Assyrian period in order to be conirmed. he paucity of pulses and fruits found in the city gate resemble low percentages observed in other areas of the site where archaeobotanical analysis is ongoing. Of those pulses and fruits represented from the city gate, there are seven lentils (Lens culinaris), one ield pea (Pisum sativum), and one ig (Ficus carica). It is worth noting that additional legume and fruit seeds have been found in other Late Assyrian contexts at the site (all in low counts), and Anatolica XLI, 2015 47 include chickpea (Cicer arietinum), grass pea (Lathyrus sativus), common vetch (Vicia sativa), and grape (Vitis vinifera). It may be said that these are all typical constituents of ancient agricultural packages from northern Mesopotamia. For now, however, the dataset of pulses and fruits remains too limited to extrapolate more detailed trends in production or consumption during the Late Assyrian period. In addition to cultivars, the occupants of Tušhan also utilized and interacted with an array of non-cultivated plant taxa, particularly, as has been stressed for the ‘Khabur Gate’ assemblage, through the vector of livestock consumption. he weeds of animal dung fuel can be linked to the environmental conditions under which they are found and lourish, and then analyzed as proxies for pastoral, farmed and irrigated landscapes. Consequently, the weed assemblage obtained from the ‘Khabur Gate’ begins the process of reconstructing the environmental mosaic accessed by the people and animals of the provincial center (Fig. 28). he predominance of grassland plants (41%) is a result of 199 seeds of timothy grass (Phleum exaratum), a popular forage among livestock, particularly for horses. he preponderance of this grass may be the result of its small size (< 1mm), and therefore its ability to avoid breakage and destruction. However, it is also plausible to see a connection between this abundant forage plant and the reliance upon horses (and other traction animals) in this frontier setting. Although the excavators of Ziyaret Tepe have not uncovered archaeological evidence for stabling, the historic record from Tušhan refers to horses as a critical component of the province’s administration in the Upper Tigris River Valley (tablet ZTT# 22 in Parpola 2008: 86-95). Accordingly, the maintenance or expansion of endemic grasslands might have been a key element of agricultural administration among Tušhan’s land managers. Steppic taxa (15%) constitute the penultimate category of weed seeds recovered and identiied from the ‘Khabur Gate’, and include may weed (Anthemis sp.), goosefoot (Chenopodium sp.), milk vetch (Astragalus sp.) and mullein (Verbascum sp.). hese are herbs and shrubs resilient in dry and sometimes saline environments, that would have been dependable graze in the foothills for livestock in the hot, summer months. he plants of the steppe are followed by ield weeds (13%) and water-loving plants indicative of wet lands (11%). Livestock consume ield weeds when they graze on the stubble and fallow from harvested ields or receive grain and straw fodder (as discussed with barley, above). hese weeds include plants historically associated with grain ields, such as bedstraw (Galium sp.), woodruf (Asperula sp.), goat grass (Aegilops sp.) and wild barley (Hordeum sp. [wild]), as well as plants that grow in a variety of habitats, but lourish on arable land, like corn gromwell (Buglossoides sp.), cow cockle (Vaccaria sp.) and campion (Silene sp.). he wet land species include purslane (Portulaca sp.), knotweed (Polygonum sp.), mint (Mentha sp.) and nut grass (Cyperus sp.). Small herds could have fed of of these plants, which would have clumped around the river’s edge and in the seepage zones of irrigation canals. To date, archaeological excavation and survey have not uncovered any ancient canal systems in or around Ziyaret Tepe, but on the basis of previous research into the extensive canal-building activities of the Late Assyrians (Bagg 2000; Dalley 2002; Ur 2005; Wilkinson et al. 2005; Wilkinson and Rayne 2010), it is reasonable to suspect that irrigation was part of crop production at Tušhan. Purslane, in particular, is a plant found in irrigated ields in Near Eastern agricultural contexts (Zohary 1966: 78), and so its abundant presence (n = 32) in the city gate assemblage is quite conspicuous. he plant can still be found around Ziyaret Tepe today, 48 Timothy Matney et al. in ields that now rely upon irrigation for crop production, and locals use the leaves as a salad vegetable. Hopefully, the continued analysis of the archaeobotanical assemblage from Ziyaret Tepe will help shed more light on this issue of water management in the provinces under Late Assyrian administration. If the representation of hydrophilic plants in Late Assyrian contexts appears particularly high when compared with weed distributions from earlier periods, then irrigation was probably a component of the agricultural regime at Tušhan. Finally, only one seed associated with woodland environments, knawel (Scleranthus uncinatus), was recovered, making up less than 1% of the weed assemblage. he paucity of woodland weeds suggests that the forests beyond the river valley had already been suiciently timbered by the Late Assyrian period, a presumption to be further tested through wood charcoal analysis. Summary: Operation Q Archaeobotany Preliminarily, the archaeobotanical assemblage from the ‘Khabur Gate’ alludes to a complex system of animal husbandry that, alongside crop husbandry, underpinned the material conditions of the Late Assyrian empire in the provincial territories. he plant remains from the city gate appear to index a range of environments utilized by animals and their human caretakers in the practice of livestock provisioning. In Late Assyrian texts, both barley and straw are taxed by the administration (Postgate 1974), and in the inscription commemorating the founding of Tušhan (882 BCE), King Ashurnasirpal II dedicates the city to the storage of barley and straw (Grayson 1991: 242-243, A.0.101.17 ii 6-28). his combined evidence suggests that foddering and foraging played an important role in the empire’s agricultural programs. herefore, although both archaeologists and Late Assyrian scribes ixate on cereal production, the irst results of archaeobotanical analysis from Tušhan also recommend that we recognize the time, energy and resources that Late Assyrian land managers put into animal husbandry as well. Postscript On-going scientiic analysis of the very large dataset generated from the 1997-2014 campaign will continue for a number of years as the inal publication volumes are being written. Many of the preliminary indings reported in our Anatolica series will require revision, modiication, or even rejection as detailed analysis proceeds. It is clear, for example, that we initially underestimated the importance of Ziyaret Tepe in the Late Roman period based on initial surface surveys. Roman remains were documented in the late 1990s with the observation during preliminary ield walking that there was an area in the western lower town with a scattering of roof tiles fragments (Matney 1998), although ceramic distribution maps showed a widespread presence of Roman ceramic types across the lower town. A concentration of Roman roof tiles, restricted in size and on the outermost lobe of the western lower town, was subsequently the object of targeted excavation in Operation J in 2002 (Matney et al. 2003: 191-194). Remains in Operation J were found to consist of walls foundations made of stone, along with roof tiles right below the surface that had been thrown up by ploughing. In Operation J there were two later phases; it remains to be established whether or not these coincide with those present in Operation T. In our earliest preliminary reports we suggested that the Roman presence on the Anatolica XLI, 2015 49 site consisted of isolated farmsteads, rebuilt once and then abandoned. Such a conclusion has now been superceded. With the discoveries in Operation T and Operation U, it is now clear that this occupation was much more extensive and evidently concentrated in the eastern part of the site. Furthermore we have, with the coins, the irst foothold in establishing a date for this occupation at Ziyaret Tepe. Details regarding the Late Assyrian city of Tušhan will, likewise, emerge through complete publication although some general trends are now well documented. For example, one overarching observation that can be made is that we see a highly conservative city plan at Tušhan that remains stable throughout the Late Assyrian period. In the Bronze Palace and Khabur Gate, the location of these major buildings remains stable across three centuries. Likewise, the general use of space in the Late Assyrian period in two phases of Operation K parallels this trend. Exceptions to this general rule exist, seen perhaps in the abandonment of the Operation Y building, and in the shift in architectural plans in Operation G/R. Of course, rebuilding and modiication of existing structures was common, but taken as a whole it would appear that the overall urban fabric of Late Assyrian Tušhan, established by royal decree in 882 BC, showed remarkable conventionality through subsequent generations before its abandonment at the end of the Assyrian empire at the end of the seventh century BC. After eighteen years of ieldwork, the closure of the Ziyaret Tepe expedition is bittersweet. An enormous dataset has been recovered and recorded and has produced signiicant scientiic results; much more remains to be done in laboratories and academic oices across North America, Europe, and Turkey. Such success is entirely due to the tireless eforts of the excellent team of archaeologists and specialists who returned, summer after summer, to labor under the hot Turkish sun. I would like to extend here my unending thanks to all my team for their eforts, energy, goodwill, and cheer, without which this undertaking would hardly have been possible. Works cited Algaze G., R. Breuninger, C. Lightfoot, and M. Rosenberg, 1991 — he Tigris-Euphrates Archaeological Reconnaissance Project: A preliminary report of the 1989-1990 season, Anatolica 17: 175-240. Bagg, A., 2000 — Irrigation in northern Mesopotamia: water for the Assyrian capitals (12th-7th centuries BC), Irrigation and Drainage Systems 14: 301-324. Behre, K.-E., 2007 — Collected seeds and fruits from herbs as prehistoric food, Vegetation History and Archaeobotany 17(1): 65-73. Cappers, R.T.J., R. Neef and R.M. Bekker, 2009 — Digital Atlas of Economic Plants. Eelde, he Netherlands: Barkhuis Publishing. Charles, M.P., 1998 — Fodder from Dung: he Recognition and Interpretation of Dung-Derived Plant Material from Archaeological Sites, Environmental Archaeology 1: 111-122. Curet, L. and W. Pestle, 2010 — Identifying High Status Foods in the Archaeological Record, Journal of Anthropological Archaeology (29): 413-431. Dalley, S., 2002 — Water management in Assyria from the ninth to the second centuries BC, ARAM 13/14: 443-460. Decker, M., 2007 — Frontier Settlement and Economy in the Byzantine East, Dumbarton Oaks Papers 61: 217-267. 50 Timothy Matney et al. Decker, M., 2009 — Tilling the Hateful Earth: Agricultural Production and Trade in the Late Antique East. Oxford: Oxford University Press. DeFrance, S., 2009 — Zooarchaeology in Complex Societies: Political Economy, Status, and Ideology, Journal of Archaeological Research (17) 2: 105-168. Grant, A., 2002 — Food, Status and Social Hierarchy. In: P. Miracle and N. Milner, eds., Consuming Passions and Patterns of Production. McDonald Institute Monographs, Cambridge. Pp. 17-23. Grayson, A.K., 1975 — Assyrian and Babylonian Chronicles. Locust Valley, NY: J.J. Augustin. Grayson, AK., 1991 — Assyrian Rulers of the Early First Millennium B.C. I (1114-859 BC). he Royal Inscriptions of Mesopotamia. Assyrian Periods Volume 2. Toronto: University of Toronto Press. Greenield, T.L., 2014 — Feeding Empires: he Political Economy of a Neo-Assyrian Provincial Capital through the Analysis of Zooarchaeological Remians. PhD thesis. Cambridge: University of Cambridge. Helbaek, H., 1961 — Late Bronze age and Byzantine crops at Beycesultan in Anatolia, Anatolian Studies 11: 77-97. Hillman, G.C., 1981 — Reconstructing crop husbandry practices from charred remains of crops. In: R. Mercer, ed., Farming Practice in British Prehistory. Edinburgh: University Press. Pp. 123-162. Hubbard, R.N.L.B. and A. Clapham, 1992 — Quantifying macroscopic plant remains, Review of Paleobotany and Palynology 73: 117-132. Jones, G., 1984 — Interpretation of archaeological plant remains: Ethnographic models from Greece. In: W. van Zeist W. A. Casparie, eds., Plants and Ancient Man: Studies in Palaeoethnobotany. Rotterdam: Balkema. Pp. 43-61. Jones, G., 1987 — A statistical approach to the archaeological identiication of crop processing, Journal of Archaeological Science 14: 311-323. Jones, G., 1990 — he application of present-day cereal processing studies to charred archaeobotanical remains, Circaea 6(2): 91-96. Jones, G. and P. Halstead, 1995 — Maslins, mixtures and monocrops: On the interpretation of archaeobotanical crop samples of heterogeneous composition, Journal of Archaeological Science 22: 103-114. Jones, M.K., 1985 — Archaeobotany beyond subsistence reconstruction. In: G. Barker and C. Gamble, eds., Beyond Domestication in Prehistoric Europe. London: Academic Press. Pp. 107-128. Kaptijn E., J. Poblome, H. Vanhaverbeke, J. Bakker and M. Waelkens, 2013 — Societal changes in the Hellenistic, Roman and early Byzantine periods. Results from the Sagalassos Territorial Archaeological Survey 2008, southwest Turkey), Anatolian Studies 63: 75-95. Marom, N., N. Raban-Gerstel, A. Mazar and G. Bar-Oz, 2009 — Backbone of Society: Evidence for Social and Economic Status of the Iron Age Population of Tel Rehov, Beth Shean Valley, Israel, Bulletin of the American Schools of Oriental Research 254: 1-22. Marston, J.M., 2011 — Archaeological markers of agricultural risk management, Journal of Anthropological Archaeology 30: 190-205. Marston, J.M. and N.F. Miller, 2014 — Intensive agriculture and land use at Roman Gordion, central Turkey, Vegetation History and Archaeobotany 23: 761-773. Martin, A.C. and W.D. Barkley, 1961 — Seed Identiication Manual. Berkeley: University of California Press. Matney, T., 1998 — Preliminary Report on the First Season of Work at Ziyaret Tepe in the Diyarbakır Province, Anatolica 24: 7-30. Matney, T. and A. Bauer, 2000 — he hird Season of Archaeological Survey at Ziyaret Tepe, Diyarbakır Province, Turkey, 1999, Anatolica 26: 119-128. Matney, T., T. Greenield, B. Hartenberger, C. Jalbrzikowski, K. Köroğlu, J. MacGinnis, A. Marsh, W. Monroe, M. Rosenzweig, K. Sauer, and D. Wicke, 2011 — Excavations at Ziyaret Tepe, Diyarbakır Province, 2009-2010 Seasons, Anatolica 37: 67-114. Matney, T., T. Greenield, B. Hartenberger, A. Keskin, K. Köroğlu, J. MacGinnis, W. Monroe, L. Rainville, M. Shepperson, T. Vorderstrasse and D. Wicke, 2009 — Excavations at Ziyaret Tepe 2007-2008, Anatolica 35: 37-84. Anatolica XLI, 2015 Matney, T., J. MacGinnis, H. McDonald, K. Nicoll, L. Rainville, M. Roaf, M.L. Smith and D. Stein (2003) Archaeological Investigations at Ziyaret Tepe, 2002, Anatolica 29: 175-221. Matney, T. and L. Rainville, editors, 2005 — Archaeological Investigations at Ziyaret Tepe, 2003-2004, Anatolica 31: 19-68. (with contributions by T. Demko, S. Kayser, K. Köroğlu, H. McDonald, J. MacGinnis, K. Nicoll, S. Parpola, M. Reimann, M. Roaf, P. Schmidt and J. Szuchman) Matney, T., L. Rainville, K. Köroğlu, A. Keskin, T. Vorderstrasse, N. Özkul Fındık, and A. Donkin, 2007 — Report on Excavations at Ziyaret Tepe, 2006 Season, Anatolica 33: 23-73. Matney, T., M. Roaf, J. MacGinnis and H. McDonald, 2002 — Archaeological Excavations at Ziyaret Tepe, 2000 and 2001, Anatolica 28: 47-89. Matney, T. and L. Somers, 1999 — he Second Season of Work at Ziyaret Tepe in the Diyarbakır Province: Preliminary Report, Anatolica 25: 203-219. Miglus, P.A., 1999 — Städtische Wohnarchitektur in Babylonien und Assyrien. Berlin. Miksicek, C.H., 1987 — Formation Processes of the archaeobotanical record, Advances in Archaeological Method and heory, Vol. 10. Pp. 211-247. Miller, N.F., 1988 — Ratios in Paleoethnobotanical Analysis. In: C.A. Hastorf and V. Popper, eds., Current Paleoethnobotany. Analytical Methods and Cultural Interpretations of Archaeological Plant Remains. Chicago: University of Chicago Press. Pp. 72-85. Miller, N.F., 1996 — Seed Eaters of the Ancient Near East: Human or Herbivore?, Current Anthropology 37(3): 521-528. Miller, N.F., 1998 — Patterns of agriculture and Land Use at medieval Gritille. In: S. Redford, ed., Archaeology of the Frontier in the Medieval Near East: Excavations at Gritille, Turkey. Archaeological Institute of America Monograph. Pp. 211-252. Miller, N.F. and T.L. Smart, 1984 — Intentional Burning of Dung as Fuel: A Mechanism for the Incorporation of Charred Seeds into the Archaeological Record, Journal of Ethnobiology 4(1): 15-28. 51 Nesbitt, M., 2006 — Identiication Guide for Near Eastern Seeds. London: University College London. Parpola, S., 2008 — “Cuneiform Texts from Ziyaret Tepe (Tušhan), 2002-2003”, State Archives of Assyria Bulletin 17: 1-113. Pearsall, D.M., 1983 — Evaluating the stability of subsistence strategies by use of paleoethnobotanical data, Journal of Ethnobiology 3(2): 121137. Pearsall, D., 2000 — Paleoethnobotany: A Handbook of Procedures. San Diego: Academic Press. Pedde, F., 2001 — Development and Extension of Near Eastern Fibulae in the Iron Age. In: R. Eichmann and H. Parzinger, eds., Migration und Kulturtransfer. Der Wandel vorder- und zentralasiatischer Kulturen im Umbruch vom 2. zum 1. vorchristlichen Jahrtausend, Berlin 1999. Pp. 485-496. Popper, V., 1988 — Selecting Quantitative Measurements in Paleoethnobotany. In: C. A. Hastorf and V. S. Popper, eds. Current Paleoethnobotany. Analytical Methods and Cultural Interpretations of Archaeological Plant Remains. Chicago. Pp. 53-71. Postgate, J.N., 1974 — Taxation and Conscription in the Assyrian Empire. Studia Pohl, Series Maior 3. Roma: Pontiical Biblical Institute. Ramsay, J., 2010 — Trade or Trash: an examination of the archaeobotanical remains from the Byzantine harbour at Caesarea Maritima, Israel, International Journal of Nautical Archaeology 39: 376-382. Ramsay, J. and A.M. Smith, 2013 — Desert agriculture at Bir Madhkur: he irst archaeobotanical evidence to support the timing and scale of agriculture during the Late Roman/Byzantine period in the hinterland of Petra, Journal of Arid Environments 99: 51-63. Reitz, E., 1987 — Vertebrate Fauna and Socioeconomic Status. In: S. Spencer-Wood, ed., Consumer Choice in Historical Archaeology. Plenum Press, London. Pp. 101-120. Rosenzweig, M., 2014 — Imperial environments: he politics of agricultural practice at Ziyaret Tepe, Turkey in the irst millennium BCE. PhD Dissertation. he University of Chicago. Schachner, A., 2003 — From the Bronze to the Iron Age: Identifying Changes in the Upper Tigris 52 Timothy Matney et al. Region. he Case of Giricano. In: B. Fischer, H. Genz, E. Jean and K. Köroğlu, eds., Identifying Changes: he Transition from Bronze to Iron Ages in Anatolia and its Neighbouring Regions. Istanbul: Türk Eskiçağ Bilimleri Enstitüsü. Pp. 151-163. Stone, E., 1987 — Nippur Neighborhoods. Chicago: Oriental Institute of the University of Chicago. Ur, J., 2005 — Sennacherib’s Northern Assyrian Canals: New Insights from Satellite Imagery and Aerial Photography, Iraq 67(1): 317-345. van der Veen, M., 1995 — he identiication of maslin crops. In: H. Kroll and R. Pasternak, eds., Res Archaeobotanicae, International Workgroup for Palaeoethnobotany; Proceedings of the Ninth Symposium, Kiel 1992. Kiel: Oetker-Voges. Pp. 335-343. van der Veen, M., 2007 — Formation processes of desiccated and carbonized plant remains – the identiication of routine practice, Journal of Archaeological Science 34(6): 968-990. Van Zeist, W., and J. Bakker-Heeres, 1985 — Archaeobotanical studies in the Levant. 4. Bronze Age sites on the north Syrian Euphrates, Palaeohistoria 27: 247-316. Wilkinson, T.J., and L. Rayne, 2010 — Hydraulic landscapes and imperial power in the Near East, Water History 2: 115-144. Wilkinson, T.J., E. Barbanes Wilkinson, J. Ur and M. Altaweel, 2005 — Landscape and Settlement in the Neo-Assyrian Empire, Bulletin of the American Schools of Oriental Research 340: 23-56. Zohary, M., 1966 — Flora Palaestina. Part One, Text. Jerusalem: Israel Academy of Sciences and Humanities. Zohary, D., M. Hopf and E. Weiss, 2012 — Domestication of Plants in the Old World: he Origin and Spread of Domesticated Plants in Southwest Asia, Europe, and the Mediterranean Basin. 4th ed. New York: Oxford University Press.