|Back to Research Summary|
|1995 Excavations at Tubna, Wadi Ziqlab|
The text of the preliminary report on these excavations follows.
Aerial View of Tubna, Map of WZ 121 architecture
Excavations at WZ 121, a Chalcolithic
site at Tubna, in Wadi Ziqlab
E. B. Banning and D. Lasby, University of Toronto, and M. Blackham, Simon Fraser University
During June and July of 1995, the Wadi Ziqlab Project carried out excavation of a Chalcolithic settlement on the western slopes of Tubna, in al-Kura district, northern Jordan. The site had been discovered during the course of a short season of test trenching in 1993.
The site lies at an elevation of about 550 m ASL in a small olive grove high on a terrace built over limestone and overlooking Wady Summayl (Wadi 'Ain Zubiya), one of Wadi Ziqlab's main tributaries. Its location is reminiscent of Hanbury-Tenison's (1987: 130) observation that in the Jerash region "Chalcolithic sites are high up towards the top of south or west facing slopes, with access to water at some distance, and covering at least two hectares." The nearest water source in antiquity, so far as we have been able to determine, would have been the stream some 100 m below the site, although it is possible that there was a nearer source in some spring that has since gone dry as a result of a fall in the water table. Only about 1 km to the south of the site, oak-pistachio forest, mainly regenerated in recent times, gives some impression of what the environment of the region may have been like before widespread deforestation. The site itself has been disturbed by agricultural activities, including plowing, tree-planting, removal of stones for terrace construction, and probably the movement of fill during terrace construction. Most of the cultural remains we encountered occurred on a single terrace, with architecture concentrated near its southern end, but probe trenches also encountered fairly abundant Chalcolithic and some Byzantine artifacts upslope, along with a single stone-lined pit of Chalcolithic age, indicating that the site may originally have extended up to where the lowest modern houses now lie. Surface lithics on the lower terrace, meanwhile, extend from just south of the excavated Areas more than 200 m to the northwest, although not all of these appear to be Chalcolithic. Many of the lithics to the far north appear to be of Middle Palaeolithic or Upper Palaeolithic age.
A test trench measuring 2 m x 1 m on this site during the summer of 1993 interesected a buried stone wall and was associated with undisturbed deposits, with abundant lithics and potsherds and fairly good preservation of bone, at a depth of approximately 50 cm. The purpose of broader excavations at this site was to determine its size, to estimate the number of structures, to uncover the plan of at least one structure, and to extract data on the character and distribution of pottery, lithics, faunal and floral remains using the same methodologies we used at WZ 200, a Late Neolithic site we had excavated in 1990 and 1992 at the confluence of Wadi Summayl and Wadi Ziqlab, about 3 km to the northwest. These include excavating by natural stratigraphic units, dry-screening excavated soils, gridding floors and surfaces into quadrats no greater than 50 cm on a side, and bagging the soil for flotation and micro-refuse analysis.
In May, before excavation commenced,
a small team carried out survey of the main terrace on which site
WZ 121 occurs. In addition to topographic survey, this included
survey by differential proton magnetometer.
The magnetic survey involves measuring differences in magnetic intensity simultaneously at a base station and at a series of points across the surface of the site, so that diurnal variations - changes in the earth's magnetic field over time - are discounted. Variations in magnetic intensity across the site can therefore be attributed to variations in the magnetic properties of materials on and beneath the surface of the earth, principally in the amount of iron present. Since the targets of greatest interest in this surface were buried walls, which we would expect to have been constructed of limestone,at least in their lower courses, we were looking mainly for linear "negative" anomalies - long, narrow areas with lower than average magnetic instensity - because limestone has extremely little iron content compared with the soil that surrounds and overlies it on the site.
Maps of magnetic variation across the site appear in figures 2 and 3. The "hot spots" with very high "spikes" of magnetic intensity mark locations where there is probably iron metal, such as nails or steel cans, close to the surface. The areas of interest are principally the "troughs" of low magnetic intensity, which we might expect to be associated with buried limestone walls. Unfortunately none of the very low anamalies is extremely linear, and there is not any really convincing evidence for buried architecture. As later excavations showed, there was indeed quite a large limestone wall that seems to correspond with a weak anomaly that runs North-South (using grid North). The most substantial negative anomalies, however, seem to mark variations in the depth to bedrock, which outcrops just east and west of the site and comes close to the modern surface in many of our excavations. A number of deep pits in Areas P21, P22 and Q22 might also have been expected to show up as magnetic anomalies; these may be the variations apparent in those squares, but they are no more pronounced than other variations in the maps that are apparently due to the sometimes extreme jaggedness of the underlying bedrock.
We imposed a 3 m x 3 m grid over the site, which we originally assumed was restricted to a single terrace on the hillside, using letters from west to east and integers from south to north, oriented to a grid North that was in fact roughly northwest. The baseline, designated "P," extends from a benchmark A on a limestone outcrop at the southern extremity of the site to the corner of a modern house visible on the hillside to the northwest. The test trench excavated in 1993, simply designated Area A, fell within the grid square now designated as N19. We began by excavating half-squares, each 3 m x 1.5 m in extent, in Areas P19, P22, P24, P27, P33, P35, N34, Q32 and Q34 to prospect for subsurface cultural deposits and architecture. Later we closed some of these Areas, especially in the more northerly part of the site, and opened additional excavation Areas in N16, M19, N19, N20, M20, M21, P21, L22, M22, N22, L23, M23, N23, M25, M26, M29, L32, L33, and M37. Many of these were placed so as to expose a reasonably large contiguous area with architecture, although the need to avoid damage to olive trees made this difficult. We also prospected for cultural remains upslope, requiring us to extend our grid farther east than we originally anticipated, with probes of either 1 m x 3 m or 1.5 m x 3 m in Areas E26, B31, X31, Z31, and A34.
Stratigraphy and Architecture of the Site
The gross stratification of the site consists of irregular bedrock, overlain by Chalcolithic deposits from 0.2 to more than 1 m thick, in turn covered by Chalcolithic walls and other Chalcolithic deposits from which pits were dug and later filled. All of this is capped by shallow deposits of Byzantine and modern age that appear to be related to agricultural activity on the site. In at least one case (in Area L32, and probably in Area N20 as well), Byzantine farmers appear to have used the remnants of prehistoric walls to pile stones during field clearance. Another Byzantine wall appears along the grid line between Areas M22 and N22. Architecture appears to be infrequent, perhaps as a result of the dispersion of households on the site, and usually not very well preserved, owing both to its proximity to the modern surface and to recent farmers' planting and stone removal.
The earliest human activity we have detected on the site appears on bedrock, and only appears in small portions in a few of our excavation Areas. Area N18 had some sherds in its earliest deposits (locus N18.008) that appeared to be Late Neolithic (particularly some splayed strap-handles), while Area P21 also exhibited some possible Late Neolithic pottery in its locus P21.008, where there seems to be some red-painted sherds. Since most of the sherds from these two loci are body sherds, and the sample size is small, confirmation of the existence of Late Neolithic materials in these contexts must await comparison of fabrics back in Toronto.
Area M19, where there may also be some earlier material just above bedrock, provides the best glimpse of what form the Late Neolithic occupation may have taken (plate I). Here a hollow in the bedrock appears to have been modified by the addition of a crude wall that closes a gap on its western side and the hollow was later filled by deposits, including some living surfaces, that contained fairly large numbers of late Neolithic or Chalcolithic sherds. It is possible that this hollow was roofed with some perishable material to provide a simple shelter. Nearby Area N18 had numerous fragments of wattle-and-daub material in locus N18.004 that could have derived from a simple structure erected at this time. In other areas of the site, especially in the north, we find abundant Chalcolithic material on bedrock, but without any features or architecture and it appears to represent either redeposited material or dispersed activity on the site.
The principal architectural features are what seem to be two very large buildings, one partially exposed in Areas L31 and L32 (plate 4) and the other extending into areas M19, N19, M20, N20, M21, N21, L22, M22, N22 and, possibly, N23 (plates 2-3), both of which must date fairly late in the Chalcolithic use of the site. In the former, we have the building's northwestern corner in Area L32, where it clearly has a foundation trench cut from the top of the preserved Chalcolithic deposits there, and much of its western side. The corner was later overlain by a Chalcolithic stone-lined pit (see below). The western wall overlies an earlier, single-leaf wall that continues into Area L31. In the latter, we have two parallel walls about 5 m apart and apparently marking the eastern and western sides of the building, but poor preservation near the surface at the north end and later pitting at the south end seem to have truncated the building's extremities and made it impossible to determine its full extent. Of cross-walls only wall M22.020 and wall A.002 (from the 1993 probe trench) are very convincing, and the former, at least, seems to be a later addition, perhaps associated with another pit like the one in L32. Neither cross-wall is bonded with the large walls. The east wall of this structure is complicated as it was built in at least two and probably three episodes. The earliest wall (N20.009) is built of extremely large stones and is parallel to wall M20.015-L22.009. When this was rebuilt as wall N20.008, it was both longer, now extending well into Areas N19 and N22, and at a slightly different angle, so that it was no longer quite parallel to wall M20.015-L22.009. As in Area L32, it seems that Byzantine farmers added stones to the ruins of wall N20.008 during field clearance, although proximity to the modern surface has removed most of these.
In both buildings construction is very similar and generally quite massive and, as we have seen, there appears in each case to have been at least one episode of rebuilding prior to Byzantine use of the walls as stone repositories. One to three courses of stones are preserved, and these are generally large, squarish boulders that give the walls something of a megalithic appearance. Most of the walls, at least in their last building phase, are double-leaf walls. We could detect no signs of foundation trenches except for the north wall in L32, but rather the walls appear to have been constructed on top of Chalcolithic surfaces, which in some cases may have been laid down intentionally to make the ground a little more level prior to construction. In spite of this, the wall on the east side of the more southerly building (N20.009) is founded at a generally higher level than that on the west (M20.015, L22.009), and the surface between slopes down to the west. The eastern wall, in its earliest phase (N20.009), appears to have extended only about 5 m from north to south, but the double-leaf wall built on top of it (N19.008-N20.008) was at least 12 m long.
We found no hearths or other domestic features within the buildings, and their exact function has yet to be determined, although it seems likely that they are portions of habitations with walled courtyards. Detailed analysis of the finds within the structures, including microrefuse from some of the surfaces in M20, N22 and N23, may help us identify activities there. Area M20, in particular, provided two good, superimposed surfaces belonging to the interior of the southern structure.
Outside the southern structure, just east of its eastern wall, we found two well constructed, stone-lined pits and a number of other less formal pits. The stone-lined pits, P22.003 and Q22.005 (plates 5 and 6), had been cut from a surface at least as high as the bottom of the modern plow zone. In fact their associated use-surfaces may well have been higher than the modern surface and have been eroded or plowed away. Both pits are approximately circular in plan, with diameters of a little under 2 m, and have nearly vertical sides lined with angular and sub-angular stones 10 to 15 cm across. The largest of the two (P22.003) was cut some 20 cm into the underlying limestone bedrock (locus 099). There may have been a third such pit in Area N18, just south of one of the large walls, where a circular floor has been chipped into the bedrock. Later pitting and rodent disturbance has obliterated traces of the upper part of this pit, if it in fact existed. Yet another similar pit occurred upslope in Area B30, but we were unable to remove the pit's fill prior to the termination of excavations.
Pits P22.003 and Q22.005 (plates
5 and 6) were filled with fine, somewhat ashy soil containing
relatively large Chalcolithic sherds and animal bones, but relatively
few lithics (pit fills P22.004 and Q22.007), and the P22 pit was
then covered with a flagstone pavement (locus P22.009) that was
only partially preserved below the plow zone. The function of
these pits, and the pavement that covered at least one of them,
has yet to be determined, but their contents seem to represent
repeated deposits of domestic rubbish, probably from hearths or
ovens, over a relatively brief period of time. The sherds in the
deposits are frequently flat-lying or gently sloping, giving the
appearance of a large number of small, interleaved surfaces, as
you might expect from the dumping of gufah-loads of soil, and
the fills' compaction varies substantially, but there is no evidence
of less ashy soil intervening between the surfaces. In Q22.007
there is a substantial disturbance, particularly in the southeast
corner from either animal burrowing or former tree roots, while
P22.004 is disturbed by the roots of a recent grape vine. The
function of the flagstones, in particular, is a puzzle. We have
other stone-capped pits at the site, generally much smaller in
size and without the ashy fill, which are also difficult to interpret.
Given what seems to have been a succession of surfaces within
the pit fills of P22.004 and Q22.007, it is possible that the
flagstones constituted a temporary cover that was periodically
relaid at a higher level as deposits accumulated in the pits.
On the upper terrace, although some trenches detected substantial deposits containing Chalcolithic artifacts, only Area B30 intersected any Chalcolithic architecture. This (B30.003) is another stone-lined pit, similar to those in P22 and Q22. Its fill (B30.004) is very rocky in its upper portion, at least, unlike those on the lower terrace. Unfortunately it was not possible to complete this pit's excavation to see if any ashy sediments occurred below the stones.
A circular, chipped depression in the bedrock in Area N19 probably represents the last traces of another of these stone-lined pits which was later destroyed by pitting and animal burrowing. Its size, shape and manufacture is reminiscent, in particular, of the bottom of pit P22.003. Another pit into bedrock in Area P24 seems to represent use of a natural depression in the bedrock. This little sinkhole was modified by chipping on some of its upper edges, but the bottom of the pit is filled with sterile Terra Rossa soil that appears to represent primary soil formation, indicating that the Chalcolithic users of the pit did not excavate to the pit's bottom.
Other enigmatic stone-lined pit features occurred in Areas L31, L32, M22 and N22. In the first three of these (plates 7, 8, 9), the features had flat slabs standing on-end around the upper edges of the pits, while a pavement of small flagstones within this ring of slabs capped the deposits within. At first we suspected that these features were small cist-graves, but excavation below the flagstones revealed only pit-fill containing Chalcolithic ceramics and no human bone occurred either above or below the flagstones. The L31 and L32 features were also lined with stones on their sides and bottoms, while the M22 feature was only partially lined on the inside and the N22 one was not lined at all. The function of these features, as with the much larger stone-lined pits, is unknown.
Finally, there were large numbers of less formal pits, some of them quite large and deep in Areas M19 and N19 and, especially, P21 (plate 10). Principal among these are the pits filled by loci 007, 006 and 003 in Area P21. As in the stone-lined pits in adjacent Area P22, some of these pit deposits were loose and ashy, and contained high frequencies of Chalcolithic pottery and animal bone, but they were also generally very rubbly, with high frequencies of angular (fire-cracked?) limestone. They also contained a fairly large number of fragments of basalt grinding implements.
The pottery from WZ 121, except for occasional sherds of Byzantine date in upper loci, consists mainly of coarsely grit-tempered pottery with parallels in Chalcolithic sites. Preliminary observations indicate that inclusions, often up to 3 mm in size, consist of chert, limestone or grog, and several distinct fabrics are recognizable. Most common are thin to medium sherds of red, yellow or orange paste with angular chert inclusions and no slip. A less common fabric is usually yellow or light brown, of medium thickness, and with a brown or buff grog or clay temper. Grey and black fabrics, again usually with angular chert inclusions, are also fairly common. A rarer fabric tends to be finer, with a thick white or buff slip and occasionally with red or brown paint. All are handmade.
The excavations yielded a wide repertoire of forms, with jars of various types apparently dominant. Many of these are somewhat inverted, without necks, making "holemouth" forms, but there are also large numbers with short everted necks and some with short to medium vertical necks that make an angle close to 90o with the shoulder. So far only two sherds, one a rim and shoulder and the other a handle and base, appear to come from churns and there are no clear examples of cornets.
Bases are usually flat and
V-shaped, sometimes with a small foot ("disk base")
and with no evidence we have been able to detect so far of mat
impressions. Ring bases and concave bases are fairly common, but
pedestal bases are very rare. There appear to be some round bases
as well, but at this stage they are difficult to distinguish with
confidence from body sherds.
Handles are mainly strap handles with relatively flat oval or somewhat U-shaped sections (raised ridges on the two edges and a trough down the middle). There is often a large, flat void inside each handle attachment that results from folding of the clay during the handles' manufacture. Smaller handles are often round or triangular in section, and there are fairly frequent examples of small to medium pierced lug handles, again with either rounded or triangular sections. Ledge handles seem to be virtually non-existent at the site.
Decoration consists primarily of applied and molded strips ("rope molding"), often arranged horizontally below the rim sor necks of jars. In one case the same sherd displayed both vertical and horizontal rope molding. Less common are jars with scalloped decoration apparently made by impressing and somewhat lifting the soft clay with a curved spatula, usually 2-3 cm in width ("spatula impressed"). Much rarer are punctate sherds, one of which has been impressed with a hollow reed. Red slip and red or brown paint occurs on a low but substantial proportion of sherds. Notably we seem to have painted decoration on the sherds with the thick white or pinkish buff slip, but so far we have not been able to discern any geometric patterns in the painting that might relate it to the "cream wares" found at other early Chalcolithic sites in the region (e.g., Kataret es-Samra [Leonard 1989], Tel Tsaf [Gophna and Sadeh 1989], Tell Abu Habil [de Contenson 1960: 35]).
Only in some of the deepest excavated contexts, right on or just above bedrock, do we find pottery that may belong to the Neolithic-Chalcolithic transition. Since the sample size from these contexts is small, and they consist mainly of body sherds, our interpretation of these contexts is uncertain. Among these sherds, however, we find wide strap handles that splay out at their attachments and with very flat sections, a handle form common in the Late Neolithic. Fabrics appear more poorly fired and softer, although they show the same kinds of inclusions as in sherds higher in the profile, and are typically brown, yellow or buff, often with a reduced (black) interior. Some of the sherds from locus P21.008, near the bottom of Area P21, seem to be red-painted, and may also belong to the Late Neolithic.
Among the formed tools, several tool types typical of the Ghassulian are common, including adzes and picks and sickle blades showing gloss but without denticulation. "Fan scrapers" are present in small numbers but appear to be made on cortical flakes from fairly flat nodules, rather than on tabular flint, which is not as commonly available in the area. Projectile points seem to be virtually absent.
The adzes and picks are almost always made on a softer, almost chalky kind of flint, quite different from the material used for the bulk of the lithics, and it seems that this was a conscious choice by the knapper, perhaps to improve the tools' resistance to fracture during use. Both kinds of material are locally available as nodules in the Upper Cretaceous limestones on which the site is found.
One of the foci of our research is an attempt to identify activity areas through study of microscopic remains of lithic debris, bone and pottery chips, shell and other materials found on house floors and outdoor surfaces. Patterns in the distributions and densities of these items, as well as statistical associations between classes of debris, may help us to identify areas where cooking, tool manufacture or repair and other activities occurred.
In three parts of the site, we gridded surfaces into quadrats of 25 cm x 25 cm and collected all the soil for this type of analysis, much as we did at the Late Neolithic site of Tabaqat al-Buma. The most promising surfaces seem to be two superimposed ones (loci 020 and 005) just east of wall 015 in Area M20, where there were numerous flat-lying sherds and stones, relatively little disturbance, and clear association with one of the large walls of the southern building. We also took control samples from non-surfaces at the site and off-site to see how well the materials from the surfaces contrast with fills and natural deposits. We will begin analysis of the new samples in Toronto in autumn.
WZ 121 in a regional perspective
Site WZ 121 joins Tell Handaquq (Mabry 1989) as one of the few excavated Chalcolithic sites in the hills of northern Jordan, and allows us to begin to see how the Chalcolithic in this area fits into the southern Levant as a whole. Compared with Ghassulian and related sites in the Negev (e.g., Perrot 1955; Levy 1987), southern Jordan Valley (Hennessy 1969; Koeppel et al 1940; Mallon et al 1934; North 1961), central Jordan (Ibrahim 1975; Hanbury-Tenison 1987) and with Chalcolithic sites in the middle Jordan Valley (e.g., Tell Abu Hamid [Dollfus et al 1988; Dollfus and Kafafi 1993]), site WZ 121 seems to have a much more dispersed settlement pattern and a much smaller repertoire of material culture. It seems likely that it represents part of a small, dispersed village or a few nearby farmsteads with a low population density and a relatively modest way of life, although fragments of two haematite maceheads and of the base of a fenestrated basalt "chalice" suggest that its inhabitants were not completely without luxuries or "elite" possessions. Although the functions of the many pits remain unclear, it seems likely that they had a domestic role, perhaps related to food-preparation activities, while the animal bones and artifact assemblage suggest that their economy was based on mixed farming with emphasis on cereals, sheep or goat, cattle and pig.
I would like to thank our associates in the Department of Antiquities of Jordan, principally Ibrahim Zu'bi, our representative, and Dr. Ghazi Bisheh, Director-General, for their assistance and support during the 1995 field season. Dr. Zeidan Kafafi kindly helped us arrange temporary accommodations for the advance team at Yarmouk University in May, and Rula Shafiq was our stand-by physical anthropologist. Abu Abdullah and Abu Khaled, the land-owners of site WZ 121, kindly allowed us to excavate in their olive grove, and numerous individuals in Tubna assisted us in various ways. Abd al-Karim Yusef Ta'ani was our capable driver and all-round assistant. Our work was supported by a grant from the Social Sciences and Humanities Research Council of Canada. The Royal Ontario Museum, through the agency of its director, Dr. John McNeill, provided support for Dan Rahimi's participation in the field season.
Our thanks go to Emerson Grossmith and Paul Racher for giving up some of their spare time to complete the photography, to Emily Rachman for maps and plans. Jo Bradley did a fantastic job running the lab, Susan Maltby as the project conservator. Mark Blackham and Paul Racher helped with site mapping and setting up camp, John Field was the project geologist, Mark Blackham was the site stratigrapher and did pH measurements, David Lasby wrote the preliminary lithics report and took care of most of our computer problems, Alicia Beck collected and recorded samples for residue analysis, Paul Racher did flotation, Rula Shafiq filled in on the field osteology, and Cindy Shobbrook was our nurse and cook. And, finally, many thanks to Area Supervisors Alicia Beck, Mark Blackham, Emerson Grossmith, Alex Hartnett, David Lasby, Paul Racher, Emily Rachman, Dan Rahimi, Percy Toop, Kathy Twiss and Ibrahim Zu'bi.
Dollfus, G; Z Kafafi; J Rewerski; N Vailland; E Coqueugniot; J Desse; and R Neef (1988). Abu Hamid, an early fourth millennium site in the Jordan Valley. Pp. 567-601 in Garrard, A. N. and H.-G. Gebel (eds.), The Prehistory of Jordan, The State of Research in 1986. BAR International Series 396(ii). Oxford: British Archaeological Reports.
Dollfus, G and Z Kafafi (1993). Recent researches at Abu Hamid. ADAJ 37: 241-255.
Elliott, C. (1978). The Ghassulian culture in Palestine: origins, influences, and abandonment. Levant 10: 37-54.
Hanbury-Tenison, J. W. (1987). Jarash Region Survey 1984. ADAJ 31: 129-157.
Hennessy, J. B. (1969). Preliminary report on a first season of excavations at Teleilat Ghassul. Levant 1: 1-24.
Ibrahim, Moawiyah M. (1975). Third season of excavations at Sahab, 1975 (preliminary report). ADAJ 20: 69-82.
Koeppel, R.; Senès, H.; Murphy, J. W.; Mahan, G. S. (1940). Teleilat Ghassul I: Compte rendu des fouilles de l'Institut biblique pontifical 1929-1932. Scripta Pontificii Instituti. Rome: Pontifical Biblical Institute.
Lee, J. R. (1973). Chalcolithic Ghassul: New Aspects and Master Typology. Ph.D diss., Hebrew University, Jerusalem.
Levy, Thomas E (1987). Shiqmim I. Studies Concerning Chalcolithic Societies in the Northern Negev Desert, Israel (1982-1984). BAR International Series 356. Oxford: British Archaeological Reports.
Mabry, John (1989). Investigations at Tell el-Handaquq, Jordan (1987-88). ADAJ 33: 59-95.
Mallon, A.; Koeppel, R.; and Neuville, R. (1934). Teleilat Ghassul I, 1929-32. Rome: Pontifical Biblical Institute.
North, R. (1961). Ghassul 1960: Excavation Report. Analecta Biblica 14. Rome: Pontifical Biblical Institute.
Perrot, J. (1955). The excavations
at Tell Abu Matar, near Beersheba. Israel Exploration Journal
5: 7-40, 73-84, 167-89.
Back to Research Summary