|
Kush.
|
Back |
 |
||
|
The site The excavations The pottery The archaeobotany The animal bones Conclusion Acknowledgements |
Kush (25° 49' 22", 56° 00' 25.7") is situated just inside the Gulf, approximately seventy kilometres south of the Straits of Hormuz. To the east, west and south the site is surrounded by the fertile and relatively well-watered Shimal plain, an alluvial deposit densely planted with date-palm groves and covered with small rural settlements. Close by to the east rise the mountains of the Musandam Peninsula, the proximity of which give the plain a high water-table. The site now lies about two and a half kilometres Southeast of the modern coast but was originally close to the edge of a lagoon which has now silted up to become a sebkha flat. The location once gave access to both agricultural and marine resources as well as trade routes, a combination which is unique on the western coast of the Oman Peninsula and which has made the Shimal area a focal point of settlement since at least the 3rd millennium B.C. The site consists of a tell measuring 120 metres north-south by 100 metres east-west. This is small by comparison with other areas of the Near East such as Iraq, Iran and the Levant where mud-brick architecture is the norm but, in the Oman Peninsula, where date-palm-frond building is the preferred type of architecture, tells do not form quickly and are quite rare. It is not a simple tell but appears to be an irregularly-shaped conglomeration of three or four separate mounds, suggesting a complicated development. The central part of the site stands 6.5 metres above the level of the surrounding plain and there appear to be at least a further 1.5 metres of archaeological stratigraphy below ground level. The site has been slightly damaged to the east and south where a cemetery wall has been constructed. Otherwise, apart from the effects of erosion, it is in good condition and does not seem to have been greatly disturbed. Trench B. In the spring of 1994 a 2-metre wide test trench was dug into the side of the mound over a four-day period. Although only a limited sounding, the excavation revealed a useful sequence. The phases uncovered are briefly described below. A This phase consists of the top few centimetres of silt and sand; somewhat disturbed but nevertheless rich in pottery and flecks of charcoal. There is some evidence of palm-frond matting within the uppermost 5 cms. These layers are quite similar in origin to those of phase ‘B’ but are more disturbed. B This phase is composed of a series of horizontally-bedded, thin, loose, silty and sandy layers, rich in pottery, bone and flecks of charcoal. The layers are very suggestive of occupation debris. Occasional postholes suggest that the occupation consisted of palm-frond huts (‘arish). C The layers which make up this phase are probably similar in origin to those of phases ‘A’ and ‘B’ but are thicker and more complicated. Together with the more or less horizontally bedded layers of loose silt and sand, are fairly thick layers of charcoal and ash interrupted by shallow cuts and hearths all suggestive of occupation debris which accumulated around ‘arish structures. Although the top three phases appear to be quite similar in make up and origin, they are separated one from the other by clear breaks in deposition. D The layers of this phase are slightly clayier and more compact than the layers of the overlying phases. This is probably due to the fact that they result, at least in part, from material eroded from mud-brick buildings. Towards the outer edge of the tell there are some lumps of mud brick suggesting the remains of walls. There is also evidence of some shallow pitting. The maximum thickness of the deposits of this phase is around one metre suggesting that it accumulated over a considerable period of time. Most of the phase is horizontally bedded. Phase D probably represents the last phase of mud-brick architecture on the site. E Phase ‘E’ is a vertically-sided pit which was cut into the layers of phase ‘D’. The lowest levels encountered were charcoal-rich. F Phase ‘F’ is made up of layers which are more gravely than those of the overlying layers. The uppermost layers are horizontally bedded whilst those lower down the slope quite steeply and appear to bank up against the sides of the structure of phase ‘G’. G Phase ‘G’ consists of a small section of a much larger structure. The part which has been exposed consists of a length of irregular but substantial walling made of rounded limestone pebbles (up to 3 cms) held together by a very compact silt and fine sand matrix which is as hard as mortar. The two metres of wall which are exposed in the sounding appear to be curved and could be part of a round structure such as a tower. The exact composition of the matrix is still to be determined. The gravel which is such an abundant component in the lower levels of phase ‘F’ must result from the weathering of this structure. H This phase consists of a circular feature constructed of dressed limestone with a small hole in the western side. It was cut during the construction of the phase ’G’ structure. I This phase is represented by a small section of gently-sloping layers consisting of compact silt and mud-brick tumble which are truncated to the west. Too little is exposed to allow an interpretation: the layers appear to run underneath phase ‘G’. J A series of steeply-sloping, silty deposits covering a thick deposit with a high humic content and green silty matrix which appears to have been a cess pit or rubbish deposit. From this layer came a large collection of Sasanian to early Islamic period pottery. Phase ‘J’ formed behind the mud-brick wall of phase ‘K’. K This phase consists of a mud-brick wall behind which Phase ‘J’ appears to have formed. It seems that ‘K’ represents a phase of mud-brick architecture which had gone out of use by the time that ‘J’ was deposited. L Phase ‘L’ consists of a series of steeply-sloping, compact silty layers which formed up against the mud-brick wall of phase ‘K’. M This phase describes a small sounding below phase ‘L’ consisting of a fine silt deposit with very little pottery underlying the wall of phase ‘K’. N Phase ‘N’ is made up of a thick, loose, silt with gravel. Interpretation of this deposit causes some problems: it may have resulted from the erosion of mud-brick buildings close by. O This phase consists of a section of mud-brick or pisè walling standing only two course high buried in phase ‘N’. P Phase ‘P’ is a similar deposit to phase ‘N’ but below wall ‘O’. Q Phase ‘Q’ consists of a small stub of mud-brick or pisè wall below phase ‘P’. This is the earliest evidence of architecture yet uncovered at the site but there is unfortunately no associated dating evidence. The final phases (‘A’ to ‘C’) appear to lack any evidence of substantial architecture. These phases can be dated to the 12th and 13th centuries and certainly appear to represent the final decline of the settlement before its complete abandonment by the 14th century. In phase ‘D’ there are strong indications that mud-brick buildings were still in use on the site although they do not appear in the trench. In phase ‘G’ there is evidence of a large and substantial structure which might be military in function which would appear to date to around the 9th or 10th centuries. Below this phase ‘J’ dates to the 6th or 7th centuries and covers phases ‘K’ to ‘Q’ in which there is evidence for mud-brick architecture on the site. There is no reliable dating evidence for these phases but they date to the Sasanian period or earlier. Trench A. In the first full season of excavation in 1995 trench A was opened. The location is intended to cut into the highest part of the tell and the aim is to excavate a complete sequence through the mound. The trench measures 10 metres wide and will extend in length from 10 metres to about 25 as excavation continues. During the 1995 season the latest levels of occupation were excavated down to an average depth of about one metre. The stratigraphy was complicated and consisted of thin occupation layers associated with numerous post-holes, ovens, a few plaster or mud floors and one cobbled surface together with a mud-brick and beachrock lined well. The entire area was very heavily pitted during the latest stages of occupation: the pit fills consisting of silty deposits containing shell, bone and pottery. There was no evidence of mud-brick architecture during the latest phase of occupation although a number of mud-brick walls began to come to light across the trench from the preceding phase. The levels uncovered are very similar to phases ‘A’ to ‘C’ of trench B and provide an opportunity to understand the latest period in the site’s life. They appear to represent sporadic occupation in ‘arish structures on a tell where the latest mud-brick structures had already gone out of use and collapsed. Interspersed with the ‘arish structures was the heavy pitting which suggests that occupation continued nearby on the tell even when there is no direct evidence for it within the area of the trench. It is hoped that the Kush sequence will allow us to tie the Iraqi, Iranian and Chinese chronologies together and solve some of the dating problems which persist in Islamic pottery studies as well as giving information about trading activities at Kush specifically and in the Indian Ocean more generally. Sherds of Celadon similar to those found by Williamson at site K103 (Old Hormuz) in the Minab Delta occur. They do not occur at New Hormuz which was founded at the beginning of the 14th century; and they are different to the vessels found on the Sinan shipwreck, dated to AD 1323, which suggests that the Kush examples should be dated to the 13th century. Also found in the surface collection were fragments of Dusun ware, probably datable to the 7th to 9th century; a sherd of Indian Red Polished Ware, which can be dated to the first five centuries AD and several sherds of Honeycomb ware whose dating is still problematic but can probably be placed in the 7th to 8th centuries AD. During the excavation of Trench A in 1995 numerous sherds of Wadi Suq pottery came to light. This material is abraded and quite obviously residual and is to be expected on a site which consists predominantly of mud-brick architecture to which large quantities of mud would have been brought, some containing residual sherds from other sites. It might, however, also indicate that a 2nd-millennium BC occupation exists beneath the mound. An Umm al-Nar ‘sugar-cube’ stone was found on top of the mound in 1994 but no other indication of occupation in this period has yet come to light. The pottery gives an impression of a continual sequence of occupation beginning somewhere between the first century and the fifth centuries AD (the possibility of a 2nd and 3rd millennium BC occupation not being entirely excluded). The late Sasanian and pre-Samarran Islamic period is well represented at the site. The ceramic chronology of this period is still problematic and it is to be hoped that the Kush sequence will make a contribution towards resolving this. There are a number of Samarran Abbasid wares at the site dating to the early to mid 9th century but, so far, the later 9th and 10th century does not appear to be well represented (e.g. early sgraffiatos and splash wares) although it is too early to be sure of this given the small area excavated. Hatched sgraffiato indicates that the site was occupied in the 11th century. However, unlike a number of sites in the Gulf, occupation continued at Kush after the 11th century, as is demonstrated by the monochrome sgraffiatos, Chinese qingbai wares and various probable Iranian products. There is certainly no occupation on the site after the beginning of the 14th century. The nearby site of al-Mataf Julfar dates to the 14th to 17th century and has been extensively excavated. None of the most common wares found there are to be found at Kush (e.g. Persian blue speckled, Khunj ware, Chinese-blue-and-white porcelain). There is some evidence of very limited activity on the site in the post-Julfar period (18th - 19th century AD) in the form of a few sherds of late Chinese- blue-and-white porcelain and fragments of locally-produced wares. Adrian Parker. To date very few systematic palaeobotanical investigations have been carried out in the Gulf. Apart from the sites at Saar, Bahrain and Failaka, Kuwait, little flotation work has been attempted. Dry sieving from a number of sites has yielded wood remains and date stones and at nearby Tell Abraq, mud brick impressions have recorded some important palaeobotanical data. At Kush a modified Siraf type flotation machine was constructed. Water was recycled between two large settling tanks below the floatation tank in order to minimise consumption. The flow of water into the floatation machine was maintained by using a 1HP electric pump. Sludging out and topping up the system with fresh water on a regular basis and the fitting of a fine filter on the intake pipe minimised the risk of contamination. Floating material was collected in 1mm and 0.3mm sieves. The heavy residues were collected in a 1mm mesh and were picked and sorted for pottery, beads, coins, glass, heavy carbonised remains, bone and shell. During the 1995 season a total of 4114 litres of soil were floated from Trench A. This yielded 4.08kg of charred material. A maximum of sixty litres of sediment per context or 100% of the context (if less than 60 litres) was floated. An initial assessment of the floatation material suggests that Kush is relatively rich in fruits and seeds when compared with other sites in the Gulf. In addition, large quantities of charcoal, numerous shells, fish bones, and smaller quantities of bone, insects and non-marine mollusca were recovered. The material recovered from the flotation was sieved into 4mm, 2mm, 1mm and 0.5mm fractions and weighed. The material was then picked under the microscope for plant remains. During the first field season modern reference plant material including wood samples was collected from different environments within the U.A.E. Preliminary results of botanical analyses. The results presented here represent around half of the material recovered during the 1995 field season. The remainder is in the process of being studied. A more informed report of the complete flotation results and their significance shall follow at a later date. This report highlights the work underway and outlines some initial findings. The material identified falls into two broad categories, those consisting of crop plants and the other non-crop plants. Crop plants. This category includes wheat, barley, lentils, dates and olives. The wheat remains were of Triticum cf. aestivem/durum type and a number of grains were found. A rachis of Triticum aestivum was also found. The material varied considerably in size and may also include ssp. compactum and spaerococcum. A number of two-rowed barley (Hordeum vulgare ssp. distichum) and six-rowed barley (Hordeum vulgare ssp. hexastichum) grains were recovered. Also found were a relatively large number of indeterminate cereal remains. Apart from these a single lentil (Lens culinaris) (of the small-seeded variety), 2 olive stones (Olea sp.) and six sidr (Zizyphus sp.) stones and charred fruits were also recovered. A large number of date stones (Phoenix dactyliferae) were encountered most of which were fragmentary, however, some were whole and measurements will be taken of these once all the material has been sorted. Non-crop plants. Within this category the following species are most likely to represent fodder crops: grass, canary grass and small-seeded legumes. A small number of grass seeds have been recovered but not as yet identified to species level. Two grains of canary grass (Phalaris sp.) have been identified. The possibility of forage plants was suggested by the presence of a number of small-seeded legumes which include Trillium. Many of these have not as yet been identified to species level as most do not have their seed coats intact. A single seed of Ranunculus type and three seeds of salam (Acacia tortilis) were also noted. Other seed types were found including four Euphorbia sp. seeds, a large quantity of (cf. Malva type ?) seeds and a number of as yet unidentified seeds and nut fragments. A quantity of wood charcoal was recovered during the flotation. Preliminary identifications include wood remains of Pheonix and Ziziphus spina-christi. Taxonomic and quantitative analyses of the carbonised wood remains shall be undertaken and the results addressed at a later date. The initial findings would suggest that, when compared with other sites in the Gulf, Kush preserves an important archaeobotanical record for the Sasanian and early Islamic period. On the whole the preservation of the material is good and the variety of seed remains found highlight the potential and importance of the sequence. Mark Beech and Alan Pipe. This preliminary catalogue of material was made on the basis of a rapid scan through the animal bones collected during the 1995 season of excavation from contexts dating to the 12th and 13th centuries AD and the 17th/18th centuries AD (excluding bones present within the flots and heavy residues, which still remain to be sorted). All bones on the excavation were retrieved utilising 3mm mesh dry sieving. The following mammals, birds, fishes and crustacea have been identified: Ovis/Capra (sheep/goat, of which more are probably goats than sheep), Bos sp. (cattle), Camelus sp. (camel), Canis/Vulpes sp. (dog/fox), Galliformes (?chicken), Chelonidae (marine turtle), Delphinidae (dolphin/porpoise), Elasmobranchii (shark/ray/skate), Sparidae (sea bream), Haemulidae: Pomadasys c.f. argyreus (silver grunter), Scaridae (parrotfish), and Crustacea: crab. The most common terrestrial mammals to be exploited were sheep and goat. The majority of those bones which could be assigned to either sheep or goat belonged to the latter. Some of the ovicaprid bones came from juvenile/immature individuals. Cattle bones were only present in small numbers. They appeared to be quite small in size and build. Four bones could be identified to camel. Two bones belonged to either dog or fox. A possible dog coprolite was discovered in context 1100 which included tiny fragments of crushed bone within its matrix. Several bird bones of galliform type (chicken?) were identified. A number of carapace fragments of marine turtle, and a single fragment of dolphin/porpoise vertebra were recorded. Out of the fishes, a preliminary examination suggested that one of the most important families were the sea breams, followed by grunts, elasmobranchs and parrotfish. Many of the fish bones showed signs of burning and several had traces of butchery marks (chop and cut marks to their surfaces). A number of crab chelae (pincer) fragments were also recovered. Some of these were burnt indicating perhaps that they had been cooked on a fire. Bone preservation appears generally to be quite good. A brief examination of the heavy residues and flots from the flotation programme revealed that there appear to be more bones within the flots than the heavy residues. The 1994 test trench and surface collection and the 1995 season have amply demonstrated that Kush has the potential to provide a good environmental (archaeobotanical and faunal) and ceramic sequence through the late pre-Islamic and Islamic periods. The work so far has also thrown some light on the development of the site. The evidence for Sasanian and early Islamic occupation is interesting. Not long ago evidence of Sasanian occupation was very scarce in the Oman Peninsula. There now appear to be a number of sites such as Khatt, Hulaylah and Kush which provide archaeological evidence for the Sasanian occupation which has been surmised from historical sources. Indeed, it seems quite possible that Kush, being apparently a Sasanian foundation, was originally one of the colonies that Wilkinson has described, and later grew into a substantial Islamic-period settlement. The pottery chronology of the late Sasanian and early Islamic periods is still problematic. Some of the material excavated from trench B seems likely to be datable to the 7th century. The chance to excavate a stratified sequence through this period may help to resolve some of the difficulties relating to the ceramic chronology. The evidence for occupation of the 9th, 10th and 11th centuries is so far limited to sherds and a few fragmentary indications from trench B. Notable is the large structure of phase ‘G’ which appears to be part of a substantial, possibly defensive, structure. More information is available for the 12th and 13th centuries - the final phase of occupation - as the excavation of trench A has already cut into layers dating to this period. It has been argued elsewhere that this period was one of economic decline in the Gulf. According to the results from trenches A and B, Kush, although still inhabited, appears to a have gone through a period of relative poverty which consisted of ‘arish structures and ‘squatter’ occupation. There is no evidence of mud-brick construction in the latest phase, although it was clearly the norm in earlier phases. Further evidence of economic decline comes from the complete lack of coins from this period, despite the fact that most of the excavated spoil from trench A was sieved through a 3mm mesh. This would appear to confirm Lowick’s observation of a lack of coins in the Gulf during 12th and 13th centuries. However, fine ceramics from Iran, China, and possibly Yemen, were imported and these enable us to date the occupation. The most common glaze wares of this period were sgraffiatos which is interesting because, from more than 40,000 sherds studied from the British excavations at the nearby al-Mataf Julfar dating to the 14th to 17th century, not a single sherd of sgraffiato has come to light. The evidence from Kush would suggest that sgraffiatos went out of circulation at the end of the 13th century; confirming what Morgan has recently pointed out in relation to the site of Old Hormuz on the other side of the Gulf. It will be interesting to compare the archaeobotanical and faunal evidence from this period with that from earlier phases once the excavation has progressed. The indications are that towards the end of the 13th century Kush was completely abandoned, probably only shortly before the site of al-Mataf Julfar was founded, and this must bring into question the present author’s conclusion that the site of early Julfar is to be located at Jazirat al-Hulaylah. No evidence of 12th or 13th century occupation was found at Hulaylah and it could be argued that Kush, being closer to al-Mataf Julfar and a much more substantial settlement, is more likely to be the site of pre-14th century Julfar. I would like to thank His Highness Sheikh Sultan bin Saqr al-Qasimi, Director of the Department of Antiquities and Museums of Ras al-Khaimah, for his kind support. The work was generously sponsored by Shell Markets (Middle East) Ltd, the National Bank of Ras al-Khaimah, the British Museum Department of Oriental Antiquities, the Gerald Averay Wainwright Near Eastern Archaeological Fund and Wolfson College, Oxford. The Shimal Folk Club lent their premises and helped in many other ways. At a personal level we owe thanks to Sheila Canby, Hala Abdul Hadi, Ajmal Khan, Firdous Khan, Robert Knox, Shanth Laxman, Mark Nesbitt, Ahmed al-Nowa, John Parsons, Abdalla Sarumi, John Slevin and Anna and Geoff Taylor. The team were: Dan Barratt, Kate Bonner, David Connolly, Murray Eiland, Katelyn Flavyn, Shahina Farid (field director), Beatrice Guiggi Morelli, Vanessa Kaoukji, Peter Moore, Adrian Parker, Adrian Powell and Ben. Firdous Khan died in 1995 and will always be deeply missed by those who knew him. |
|
© National Museum &
|