ÇATALHÖYÜK 1998 ARCHIVE REPORT

Excavations on the West Mound at Çatalhöyük 1998

Jonathan Last

During July and August 1998, a small team of British and Turkish archaeologists conducted excavations on the West Mound at Çatalhöyük to assess the nature and condition of the archaeological remains there. This work was carried out with the aid of a grant from the Wainwright fund. 

Background & Aims

In 1961 James Mellaart opened two trenches, on the top of the mound (Trench I) and on a low plateau in the south-west (Trench II). Large amounts of pottery of Early Chalcolithic date were recovered (90% of all artefacts), which Mellaart divided into two phases (I and II) on the basis of formal and decorative attributes, the second phase represented by a series of pits in Trench II. No evidence for coherent architecture was found in Trench I, while mudbrick walls and internal 'buttresses' of a poorly preserved structure (of Early Chalcolithic I) were revealed in Trench II. The publication of this work (Mellaart 1965) focusses on ceramic typology, with little discussion of other artefacts, site stratigraphy or context, and a sketch plan and section only for Trench II. The small scale of the interventions and the cursory nature of the report ensured that many questions about the nature of occupation on the West Mound remain unanswered. 

Surface scraping in 1994 as part of the renewed work at Çatalhöyük apparently confirmed the lack of Chalcolithic architecture close to the surface; instead the scraping revealed brick-lined graves (probably Byzantine) cut into pre-existing plough-disturbed deposits (Matthews 1994). Ploughmarks post-dating these graves confirmed the ongoing agricultural alteration of the site, which reflects its lower, flatter topography compared with the Neolithic East Mound, where such disturbances were not encountered. As in Mellaart's soundings, however, abundant Chalcolithic pottery was again discovered during the scraping. This material, as well as that collected from a systematic surface survey of the site, adds typologically to Mellaart's assemblage but provides no new contextual information other than broad differences between the uppermost deposits in different parts of the site (Last 1997). Little new evidence has been acquired about the nature of the site, though geomorphological coring has demonstrated that archaeological deposits extend even more deeply below modern ground level (which has risen several metres since prehistoric times due to alluviation) than they do under the East Mound. 

The current Çatalhöyük Project also includes a regional survey of sites in the Konya Plain. This work has revealed at least five more sites in the region with Early Chalcolithic pottery, and preliminary analysis suggests differences in fabrics, forms and finishes between Çatalhöyük West and the most important 'new' site at Çinili Höyük (Gibson 1997). 

Further observation at Çatalhöyük in 1997 found that renewed ploughing on the south side of the West Mound was bringing large quantities of pottery to the surface. Because of the difference in colour between the soil of the mound and the surrounding alluvium the extent of ploughing could be estimated at 100m up onto the mound. Despite the long history of agricultural activity on the site the possibility of further damage to intact archaeological deposits seemed high and, given the importance of the site, the extent of this threat required evaluation. 

The West Mound Project was therefore initiated with a series of aims in mind: 

(1) - In the light of the ambiguous results of the work conducted in 1961 and 1994, to evaluate the nature of archaeological deposits, the preservation of the site and the formation processes which have affected it, including the extent of damage from recent agricultural activity. 

(2) - To apply new techniques to the Chalcolithic deposits (e.g. flotation, micromorphology) in order to provide data comparable to those from the Neolithic East Mound and contribute to the longer term picture of settlement at Çatalhöyük. 

(3) - To recover a stratified ceramic sequence, as well as other material culture and environmental assemblages, in the light of questions about chronological and cultural continuity between the East and West Mounds at Çatalhöyük (Mellaart 1965; Last 1997). 

(4) - To reassess Mellaart's ceramic typology in the light of recent discoveries, including the ongoing work of the Konya Plain Survey. 

(5) - By comparison with sites such as Can Hasan (French 1998), to contribute to answering broader questions about the origins and nature of the Chalcolithic in the Konya Plain and Anatolia generally. 

Methodology and Results

Trench I

The northern section revealed a relatively thin topsoil (c 0.2m thick), overlying mixed deposit 2983, a dark greyish brown silty clay loam c 1.2m thick. By analogy with the 1994 scraping this was considered to be a Roman ploughsoil. However, plastery and ashy lenses, interrupted by numerous animal burrows, were visible up to 0.4m below the surface, generally sloping down from west to east. In the lower part of the deposit, below the step in the section, 2983 contained numerous fragments of charcoal, mortar and brick, as well as some large Chalcolithic potsherds, and this should be considered an intact - if disturbed - prehistoric fill or make-up. At the east end of the trench, the walls discussed below reached a height of 1007.46m, 0.75m below ground level, which represents the lowest level that could have been reached by the plough in antiquity. The sides of a rectangular pit (2984), perhaps a grave, 2.7m wide and 0.95m deep, which cut 2983 could be traced right up to the base of the modern topsoil. It contained in its base fragments of large tiles, which were often used to cover or line Byzantine burials, but here appeared to have been partly disturbed by a later cut or re-cut. At the east end of the step was another grave, in which human remains were discovered but not excavated. This remained unnumbered because its cut was not clear in section, having been disturbed by an irregular feature of more recent origin (2986). 

In general, therefore, there is little evidence for substantial agricultural reworking of archaeological deposits on the top of the mound, but Byzantine pitting and the ubiquitous soil animals have made a significant impact in this area. Other cut features were visible on the upper step of the trench, including the base of a post-hole (2905) about 0.25m in diameter, with an ashy fill (2904). This contained a single Chalcolithic sherd, but the feature may well be of a later date. It has some bearing on the interpretation of a small pit (2935) revealed beneath 2905 when cutting back the section below the step. This measured 0.55 x 0.4 m and contained in its base a number of unworked limestone blocks, deposited with a large part of a human skull, as well as fragments of one or two others. Morphologically the robust build of these skulls is more akin to the Byzantine population of Çatalhöyük than the Neolithic one. However, no definitely Chalcolithic human remains have been found at the site so a direct comparison is not possible. Moreover, the practice of depositing skulls separately from other body parts seems more likely to be prehistoric than Classical. Because the feature was very close to the side of Mellaart's trench, the stratigraphic evidence is indecisive. It lies between two large Classical pits, 2984, which was mentioned above, and a large circular cut about 2m in diameter just to the south (2919), which was revealed in the base of Mellaart's trench. If Chalcolithic, the skull pit survived rather fortuitously; if Classical, it must have been a deep shaft or post-hole in an area of burials or other pits, itself cut by 2905. None of the pottery from the fills (2914, 2922) of 2935 is definitely Roman, while mean sherd size is greater than in the fill of 2919. 

The latter feature dominated the area exposed in the western part of the trench. Its fill (2918), not fully excavated, contained large amounts of broken-up Chalcolithic pottery, as well as a few late items such as iron (?coffin) nails. Of uncertain date are some disarticulated human remains, again from a robust individual, and numerous small fragments of hard white lime plaster, with traces of red painted designs. The plaster may, like the pottery, be redeposited prehistoric material, but it occurred only in this late feature and does not resemble the Chalcolithic plasters from other deposits. Its origins remain uncertain and it awaits further study. 

Another late feature on this upper step of the trench base was the western end of a rectangular cut (2936), 1.5m wide, which was lined with broken tiles. Although no human remains were found, it is almost certainly another Byzantine grave, most of which was removed during Mellaart's excavation. Both 2919 and 2936 were cut through Chalcolithic deposits consisting of a compact, brown bricky fill interspersed by lenses of ash and charcoal, sometimes containing fragments of white plaster. The deposits sloped down from north to south. They were sampled at this level (2923) and seen also in the north-south section marking the step in the trench base (cut back as deposits 2953, 2932, 2937, 2938, 2954, 2955, 2957 and renumbered more precisely in section). This section shows a series of layers and lenses of varying thickness and make-up, not as rich in organics as the 'midden' deposits familiar in the Mellaart area of the East Mound but apparently representing a lengthy series of slopewash, destruction deposits, discard events and probable casual surfaces. Some of the layers contain sizeable chunks of mudbrick and mortar, like the building fill discussed below. Others have a relatively high density of potsherds lying flat as if discarded on a surface. That there were working surfaces in this area is suggested by the remains of the possible fire installation 2962, also discussed below. Many of the deposits have a fine laminar structure, and micromorphological analysis should determine whether this reflects a natural or anthropogenic origin. 

The most interesting features appeared in the lower, eastern part of the trench. According to Mellaart (1965, 135) there were no structures in Trench I, but we discovered walls preserved at different levels in the section and base of the trench, suggesting they had been truncated by the original excavators, who had failed to recognize them. We revealed the south-western corner of a building of Early Chalcolithic I date, constructed, just like the Neolithic houses, from double walls of plastered mud-brick. Unlike the Neolithic structures, however, the building appears, on this side at least, to have been free-standing, rather than abutting other buildings. In the scraped area on the northern part of the East Mound, walls abutting external spaces, such as the alley or street revealed there, are generally single. There is no detailed information about the construction of the walls of Mellaart's house in Trench II. 

As revealed, the western wall of the Trench I building runs 1.8m and the southern wall 2.9m. 1.8m east of the corner, another wall (F304) runs back into the northern section. Because of the positioning of the trench this was not completely revealed, so it is not clear whether it is a double wall or how far the southern wall continues beyond it. However, wall 304 clearly abuts the inner southern wall (F303) and it therefore seems likely that the former is an internal feature, delimiting a narrow western room within a larger house (not unlike Space 70 in B.1 on the East Mound). Otherwise the building would be a very curious shape, and since there are no apparent constraints of space on the south or west sides, this seems rather unlikely. Certainly the building revealed in Trench II, though not as regular as the Can Hasan buildings (Mellaart 1965, 136), was apparently rectangular (c 9 x 8m) with internal buttresses and platforms projecting up to 2.25m into the house; wall 304, which does not necessarily extend more than 1.3m, may be a similar partition. 

The room was excavated down to its latest floor, and I describe the deposits in chronological sequence. The walls were constructed of dark greyish brown silty clay mudbricks measuring c 0.8 x 0.4 x 0.1m, interspersed with deposits of paler mortar 30-40mm thick. Neither the bricks nor the mortar have significant inclusions. The brick size appears very similar to that recorded at Can Hasan in layers 3/2B (French 1968). If the abutments of the walls reveal the construction sequence, then the order ran: outer west (F300), outer south (F302), inner south (F303), inner west (F301). It is unclear whether the double walls represent the original construction or a later rebuild, perhaps to shore up failing walls. In favour of the former hypothesis is the lack of significant plaster layers on the internal faces of the outer walls (though these could have been stripped off), and the similar constructions of both F300 and F301, which show a very even and regular build up to about 0.7m above the floor, and more haphazard coursing above that (see section on Figure 25). F304 shows a similar construction, with the upper brickwork apparently leaning into the room. The truncated southern walls were preserved only some 0.1m above the latest floor and little may be said about their construction, other than the lack of any plaster or mortar rendering on the outer face, despite this presumably having been exposed to the elements. The inner faces of F301, 303 and 304, or at least their lower, neater courses, were all rendered with a greenish plaster 1-2cm thick, which appears to match the 'greenish-yellow plaster' described by Mellaart (1965, 136), and is very different from the white, laminated marl plaster of the Neolithic houses on the East Mound. 

In general, however, constructional features such as brick sizes seem to correspond very closely to Neolithic practices. Further analysis of the mudbricks is awaited, but sherds recovered from the bricks of F303 were found to include two probable Neolithic fragments, apparently of level VII type. This may well imply that the Chalcolithic population were using the abandoned East Mound as a source of clay for mudbricks; such re-use may help explain why so much Chalcolithic pottery has been found on the East, without corresponding evidence for structures (J. Mellaart, pers. comm.). 

The latest floor in the building (2971) was found at a height of c 1006.1m, only some 0.1m below the tops of the walls exposed in the base of the trench, but up to 1.35m below their full height as visible in the northern section. A total area of 2.2 x 1.1m was exposed, representing the south-western corner of the building. The floor, which has been sampled for phytoliths and micromorphological analysis, is of beaten earth rather than laid plaster, and rather messy in that it contains fragments of plaster, brick and animal bone within its matrix. However, it clearly represents a proper surface, from its consistency, evenness and compaction; and not least because sitting upon it were two grindstones (2971.X1, X2), which have been sampled for residue analysis. These were the only macro-artefacts upon the floor, although the immediately overlying deposit was quadranted and sampled for flotation (Units 2966-2969). A few sherds came from this layer; interestingly three of the four rims were plain slipped or burnished, while the fourth was from a rather curious, probably miniature pot in an unusual chaff-tempered fabric. Of 36 body sherds, only 11 were definitely painted (30%, where the average for the site is almost 50%). Does this suggest an unusual assemblage in this area of the house, or is the material just part of the building fill above, from which it was not stratigraphically distinct? 

The overlying deposit (2960) is a friable, dark greyish brown silty clay loam, with a high density of building debris: plaster, mortar and brick fragments, one of the latter (2960 sample 7) as big as half a brick. The deposit represents a rapid accumulation, probably part of a deliberate infilling, topped as it was by a formal deposit (see below). Some tip-lines suggest the direction of filling was from the east, i.e. the interior of the building, although too little was excavated to make much of this. 2960 contained a typical range of pottery, again with a rather low percentage of decorated sherds but six painted rims were found, generally of small vessels. These include two small S-profiled bowls with unusual decoration (Figure 26.1-2): one has a checkerboard pattern on its exterior like a piece from Mersin level XXIV on a vessel of different type (Garstang 1953, fig. 36: 32; cf. Mellaart 1965, fig. 6.4, where similar decoration occurs on the inside of a shallow bowl); the other has on both surfaces patterns like those on the inside of Mellaart's carinated bowls (e.g. ibid., fig. 5.7), but sinuous rather than jagged and without a lower border. 

At a height of 0.9m above the floor, a clear boundary was visible between 2960 and the overlying fill 2958; the latter differed from 2960 in its inclusions, containing numerous fragments of red-painted plaster compared to the white plaster and beige or red-brown mortar in the fill below. This plaster was also distinct in composition from that in 2918 and appears to have an overall coat of paint rather than particular designs. It is interesting that the inclusions not only differ between the two phases of the building fill but also appear qualitatively distinct from the plaster on the adjacent walls. Possibly this reflects the sequential collapse or demolition of different parts of the building, e.g. roofing material, then the upper parts of the walls. At Can Hasan there was evidence in some of the houses for upper storeys, from which painted plaster had fallen (French 1962), and the lack of painted plaster on the standing walls may suggest a similar situation at Çatalhöyük. Whether the building was filled by pushing in the superstructure of the house or with material transported from elsewhere, the presence of these different plasters in distinct deposits suggests a range of construction techniques in the Chalcolithic, beyond what has been observed in the small areas excavated so far. 

The pottery in 2958 was generally more broken than that from 2960 but includes slightly more decorated pieces, with a few more unusual designs as well as a small crescent lug on a painted body sherd (Figure 26.3; Mellaart shows these only as large examples on coarseware jars, just below the rim). Hence the material from within the building fill, though comparatively scanty compared with the densities of pottery in extra-mural deposits like 2923 and 2939, immediately adds something to Mellaart's typology for Early Chalcolithic I. 

The western wall of the building had apparently been deliberately levelled off at a height some 1.3m above the floor. In the top of 2958 (the upper spit of which was removed as 2951) and lipping over the top of the levelled-off wall F301 was an unusual group of items which apparently comprises a ritual deposit perhaps marking the closure of the building (2952). The objects include twelve objects which can loosely be termed 'potstands', two miniature unfired clay vessels, some animal bone fragments, and a large sherd of a pottery vessel with an unusual grey burnished surface. From the layer above (2943), the fill of an irregular pit some 0.6m deep apparently including material disturbed from this deposit, were a further six 'potstands' and a lot more of the burnished pot, as well as a complete small lugged bowl. The initial cleaning and emptying of the trench (2902, 2920, 2942) had produced another eight 'potstands' from the vicinity. As a whole, these represent the most significant group of objects found at Çatalhöyük West and so far a unique assemblage. 

Potstands are relatively common at Çatalhöyük, particularly in Early Chalcolithic II, but all those recorded by Mellaart are large, coarse and decorated with incised geometric designs (Mellaart 1965, Fig. 10). None of this type were included in deposit 2952; they were all much smaller, and the majority of them have two 'horns' rather than the single prong seen on all the examples illustrated by Mellaart. Only five of the 26 potstands have a single prong and they recall the Neolithic potstands from a level III house on the East Mound, now in Ankara Museum, rather than the large incised Chalcolithic examples. Three of them (2943.D1, 2952.X9, 2952.X20) have a pinkish fabric with some calcareous grits, essentially the same as one of the main pottery fabrics, a cream slip over this, and a relatively square base (c 35x32mm). They range in height from 55 to 60mm. The fourth (2902.D6) is poorly fired in a brown fabric with dark grey surfaces; it is of a similar shape but c 80mm tall. There is another similar fragment from 2920. 

The remainder of the 'potstands' all have two prongs or horns (Figure 27); they vary in size and form but tend to occur in pairs of similar type. None of them are slipped but many have smoothed or wiped surfaces with fine striations. Four are dark grey, varying in fabric, but with rectangular bases; they range in height from 71 to 83mm with prongs 10-20mm apart. These appear to form two pairs, 2943.D2 and 2952.X6 being very similar while 2902.D4 and D5 are more divergent in form but generally coarser in fabric. A fifth example (2942.D1) is similar, 70mm high but unfired and incomplete. Two more (2902.D1, 2943.D3) are also similar in size (80-82mm high, 9-13mm between prongs) but have squarer bases and light brown or orange surfaces. An additional fragment appears to be from another unfired potstand of similar size and fabric (2943.D5). A further two are of medium size (65-67mm high) in a similar colour and with a similar profile, both having relatively long prongs, but they differ in some features: one has a square hole through the body, as on some of the large incised potstands (Mellaart 1965, Fig. 10), and parallel prongs, 8mm apart (2952.X10); the other has no hole and diverging prongs up to 15mm apart (2943.D4). 

The other eleven objects are very different. Two are miniatures of the general form described above (2952.X12; the other from 2951 sample 2); they have a grey fabric, 22-24mm high with circular bases 15-18mm in diameter and small prongs. The remaining nine have widely spaced 'horns', rather than prongs, and do not closely resemble the standard 'potstand' form, though they are clearly derived from it to the extent of having two projections from a solid body. The two closest to the potstand form are of medium size (53-55mm high, base c 53 x 27mm) with forward-pointing parallel prongs, c 20mm apart 2952.X17, X18). One of these is broken. Two are slightly taller (61-65mm) with upright horns 28-37mm apart (2952.X1, X11). Two more (2902.D3, 2952.X5) are larger again (86mm high; base area 80 x 32-39mm) with orange surfaces and very widely-spaced upright horns c 50mm apart (these are broken off in one case). Another pair are not tall (51-53mm) but have very widely spaced horns 52-54mm apart (2952.X8, the other from 2943 sample 2). The final object, which came from 2902, seems to be a fragment of a large horned object, the base of which measures c120 x 60mm, with the horns broken off. 

The specific forms of these objects are not random: for instance 2952.X5 and X11 have very differently shaped horns, while 2952.X1 has one horn of each type. Nor does the pairing of many of these objects seem accidental; the unfired pots mentioned above (2952.X7, X14) also form a deliberately placed pair either side of 2952.X6 - one inverted, the other the right way up. That the selection and placement of objects was crucial to the deposit is shown by the positioning of the proximal end of a sheep femur between the prongs of one of the potstands (2952.X1; it was this pair of objects that lipped over the top of the wall, which therefore also looks like a deliberate placement). A number of other significant symbolic dimensions can be adduced, producing a table of oppositions which intersects with the pairing of forms: 

The unique 'horned' form of these objects clearly recalls the bucrania of the Neolithic houses, but may also point towards objects like the 'horns of consecration' from an EBA shrine at Beycesultan. It is interesting that the Level III (Late Neolithic) deposits at Kösk Höyük near Nigde included 'pot supports in the form of stylised bull's heads' (Yakar 1991: 192). The symbolic importance of potstands is attested also at Can Hasan, where anthropomorphic potstands were found in levels 2B and 2A, with horned forms appearing in the Late Chalcolithic (French 1965; 1967). Perhaps in this domestic context and with the emphasis on the symbolic construction of oppositions and analogies, they mark a liminal point between the 'death' of one house and the 'birth' of another. Although the artefacts were apparently deliberately placed within fill 2951/2958 on and next to wall F301, so many objects came from pit-fill 2943 (cut 2949, 2.1m wide and 0.5m deep), that this might be a related feature excavated in order to bury them, or perhaps to alter or recover part of the deposit. There was no evidence for Classical material in the pit, while the shape of the cut certainly seems to show a knowledge of the wall-lines of the building. 

Of the other finds within 2952, the unfired pots both appear to be flat-based with a cylindrical profile and around 50mm in diameter. The vessel comprising sherd 2952.X19 and pieces from 2943 and 2950 is a well-made carinated bowl, oval in shape, with a ring-base and a small knob on the carination. The surfaces are burnished light to dark grey producing a rather 'metallic' finish. All of the base and about 1/3 of the rim and body have been recovered. Overall, it seems very similar to one illustrated by Mellaart (1965, fig.5:3). 

Unit 2950, the deposit above 2951 and cut by 2949, appears to be related to 2952 by virtue of the presence of another small piece of this vessel. A further unusual sherd from 2950 is a painted body with alternating filled and empty triangles below the neck zone (cf. Mellaart 1965, fig. 3:2), although here the paint within the triangle is a different colour to that of the lines. As well as the 'potstands', pit-fill 2943 was also notable for the complete base of a large vessel, presumably a globular jar as in Mellaart's (ibid.) figs. 2-3 (none of which reconstructions include the bases). The underside of this base shows possible concentric mat impressions, possibly indicating the use of a tournette during manufacture. There was also a hole in the vessel wall, clearly drilled after firing and perhaps therefore a mend-hole. This vessel comprises considerably larger sherds than anything else from 2943 so perhaps it too was related to the ritual deposit (from which it was not greatly distant). Compared with the rather unusual assemblages in 2958 and 2960, 2943 includes more decorated sherds and a range of standard forms, but it also produced a carinated bowl with broad vertical bands of paint on the interior, part of a small painted handle, and a small, coarse ware lugged bowl, with a slightly more sinuous profile than Mellaart's (ibid.) fig. 11:6. 

Outside the building a 0.1m spit was excavated on the south side of wall F302. The matrix in this extra-mural area (2939) comprises a friable, dark greyish brown silty loam, with fewer lenses and layers than the deposits to the west of the building (see above). It contained less material than 2923, particularly pot and bone, while the sherds are generally smaller than in 2923 but larger than those from the building fill. 2939 also contained a low density of plant remains, compared both with the East Mound generally and with building fill 2958. Perhaps the animal bone gives a clue to the different origins of this deposit compared to 2958: while the latter contains all body parts, albeit highly fragmented, deposit 2939 produced largely post-consumption remains. In terms of pottery, 2939 had a number of rim-sherds of typical Early Chalcolithic I painted vessels, mostly bowls, as well as a number of small knobs or lugs on plain ware vessels. 2923, in contrast, had a greater range of forms, with rather more jar necks and ring-bases (from carinated bowls), though the numbers are small in both cases. A single dark painted sherd from 2939 looks more like an Early Chalcolithic II type. An unusual painted sherd with dense chaff temper may be an import - chaff seems to be less common as a filler at Çatalhöyük than elsewhere on the Konya Plain (particularly Can Hasan). 

Against the exterior of the west wall of the building, at a height c 0.1m above the surface inside, we found the destroyed remnants of a possible fire installation (2962). The deposit contained numerous large fragments of dark grey burnt clay, smooth and lighter in colour on one surface, and including a considerable density of vegetable matter. This material recalls the 'oven lining' associated with better-preserved fire installations on the East Mound. Although most of these fragments are not necessarily in situ, some of them appear to run into the bricks of the wall, which suggests a structure cut into the wall - a feature also seen on the East Mound, though only for ovens within buildings. 

Trench II

Whatever the original level and extent of the floor of this building, it seems that the deposits revealed in the base of Trench II represent the sub-floor fill or make-up. At a height of c 1002.80m a number of deposits were exposed in plan, without clear boundaries between them. In the western half of the trench, 2963 was a disturbed, cloddy, greyish brown clayey silt, with patches of soft, ashy material. It was excavated in a 1 x 1m test-pit in the south-west corner of the trench. This deposit perhaps includes the remains of surfaces, marked in one case by several sherds of a large pot lying flat at a height of 1002.73m. However, the section was too root-disturbed to show any layering. The ceramic material from 2963 lacks the dark painted wares of the Phase II pit-fills, but includes a jar neck with parallel horizontal bands of paint, a trait of Phase II, and a plain holemouth jar with a large crescentic lug: although Mellaart puts these into both his ceramic phases, we found them otherwise only in Phase II contexts. It may therefore be that 2963 - and the rest of the building - represents a later period within Phase I than the Trench I deposits, which might also explain the Phase II-like material in 2978. 

Below 2963, deposit 2964 was arbitrarily defined at the level where root disturbance became less severe, c 0.2m below the base of the trench. This was a massive deposit, essentially of the same material as 2963. It produced an extremely high density of plant remains, and quantities of pottery comparable to the extra-mural deposits 2923 in Trench I (considerably more than the Trench I building fills). It contained more red-painted sherds than 2963, and more typical Phase I shapes, including ring-bases and carinated bowls with small lugs. Some time gap between these deposits and those excavated in Trench I, however, is still indicated by differences in fabrics and slip colours. Unusual designs include a concentric diamond motif, perhaps from a carinated bowl (Figure 26.6). 

At the base of 2964 was a deposit of hackberry seeds, which resembled those from the midden layers on the East Mound. This lay above a deposit of white plastery material (2976), sloping down from the south-west at a height of 1002.53m, and partly overlying in turn a layer of soft ashy fill (2977: 1002.38m). These were not excavated and given their small exposure they are hard to interpret structurally. Overall, however, the character of deposit 2964 and the underlying layers resemble the extra-mural deposits in Trench I more than the building fill. 

Deposit 2963 ran out towards the east of the trench, where Mellaart's 'buttress' was left standing. On the north side of this, the deposit (unnumbered and unexcavated, but seen in the side of pit 2941) was a massive, compact greyish clay. To the south of the buttress, deposit 2980 was a soft, mixed plastery fill with a number of large sherds, not all lying flat. This seems to be fill around 2979, a rectilinear feature of clean, massive and compact silty clay, c 1m wide, which was not excavated but possibly represents the top of an earlier wall or buttress on a different alignment - it appears to be an upstanding feature surrounded by a softer fill, rather than a cut. 

The buttress (2965) was partly cut back in the course of cleaning up a section. As preserved, it consists of six courses of dark greyish brown mudbrick, each only about 5cm high, bonded with thin layers of a lighter, sandier mortar (rather different from that in the Trench I walls). It appears to sit upon deposit 2980. The brickwork seems to have subsided somewhat, and now slopes down rather steeply to the north. The feature was originally perhaps 0.8m wide, approximately the length of the mudbricks used in the walls found in Trench I, or twice the width - though there is no sign of a double wall here. 

In summary, although the deposits revealed in the base of Trench II are rather ambiguous, partly because of their poor condition and partly because of their limited exposure and excavation, the pottery from them appears to be of Early Chalcolithic I type, though perhaps of a later sub-phase than the Trench I material. Future analysis will consider the extent of these differences. 

Although the Phase I layers provide interesting comparisons for the more extensively excavated deposits in Trench I, the work in Trench II concentrated on the Early Chalcolithic pits, which were seen mainly in the sections and the steps against them. There is a distinction between those with stratified fills and those with homogeneous, generally softer fills. The layered deposit 2978, discussed above, seems to lie within the earliest cut evidenced. This was overlain by a massive deposit of compact greyish-brown silty clay (2982), up to 0.9m thick and dense enough to appear whitish in both plan and section - where it was not riddled with animal burrows. Presumably this represents a building fill, with a few lenses of orange daub or clay in its upper part. It was cut by a series of features, notably in the east by the large pit 2941, which is discussed in detail below. On the west side it was also cut by 3428, which is over 2m in diameter and up to 1m deep. This was filled by 2981, a generally homogeneous soft, dark greyish brown silt loam. A few large sherds were recovered from the section of this feature. 

Higher up, 2941 was in turn cut away on its western edge, by 2945, which is almost certainly the eastern half of the northernmost pit labelled on Mellaart's plan (1965, Fig. 1). Its original dimensions were probably c 1.75 x 1.5m. Its fill (2944) consisted of a relatively soft, dark greyish brown silty clay, without the tip-lines and lenses seen in the fills of 2941. As excavated it was up to 0.3m deep and contained a large number of sherds, though these were generally smaller in size than those from 2941, and at a lower density than all but the lowest fill of that feature. In the top of the northern section, another cut was visible with a very soft greyish fill (unnumbered); this is probably a recent feature, akin to the similarly loose 2987 in Trench I. 

But it was 2941 which dominated the trench, and excavation concentrated on sectioning this large pit of Phase II, which Mellaart's trench would barely have clipped. The feature is over 2.25m deep (it was not bottomed) and apparently several metres across. Extrapolating the arc of the western edge into a circular feature gives a diameter of c 5m. The pit was filled with a series of dump layers and lenses which contained a great quantity of potsherds (totalling over 134kg), often forming clear tip-lines in section, as well as much other artefactual and environmental evidence. The fill was divided into four semi-arbitrary units. The lowest of these, 2961, was a deposit of relatively clean, compact greyish-brown clay up to 0.8m thick around the sides and (presumably) base of the pit. It was more massive than the fills above and resembled the material through which the feature was cut (2912), except for being slightly more mixed in composition, with occasional lenses of plaster or ash and sparse fragments of daub. It also contained a number of sherds lying upright against the sides of the cut. It may represent a partial backfilling using the same material that had been excavated from the feature. 2961 contained a large quantity of Early Chalcolithic II pottery and other finds, though at a lower density than the overlying fills. 2959, above, was a dark greyish brown silty clay, less homogeneous than 2961 but firmer and greasier than the deposits above. As excavated it was 0.45m thick. The density of ceramics and stone from two samples in 2959 was greater than anywhere else on the West Mound except the lowest layer within 2910 above. 2911 and 2910 together represent a division into two phases of perhaps a dozen major depositional episodes, with many finer layers and lenses within these. The soil matrix is generally a dark greyish brown silt or silty clay, but the density, size and nature of the inclusions and lenses vary from massive deposits with large lumps of daub or burnt clay, to finely lensed, charcoal-rich midden-like layers; the division between 2910 and 2911 represented an attempt to follow two major lenses consisting of soft grey ash overlain by a fine yellowish sand, which sloped steeply down within the pit-fill. Flotation samples from the different components of 2911 show a much greater quantity of pot, bone and stone in the midden-like layers than the ashy/sandy lenses; the pottery in the latter was also considerably more broken. Similar differences are apparent between the samples taken from the top and bottom layers of 2910, the latter producing more of everything except obsidian, and having a mean sherd size twice that of the upper sample. The organic content of the layers also clearly varied, with chemical processes leaving a greenish/orange staining on much of the pottery, which has sometimes attacked the iron-rich paint more than the slip. This is most pronounced in 2959, significant in 2961 and 2911, and rather less common in 2910. 

Despite the variety of depositional events evidenced, the pottery from 2941 forms a rather uniform assemblage of Early Chalcolithic II, with larger vessels on the whole than Phase I and in particular an increase in necked jars, often painted with horizontal bands (Figure 26.4) and sometimes provided with horizontal strap handles, globular bowls with restricted mouths (Figure 26.5), and holemouths at the expense of the carinated and S-profiled bowls of the earlier phase. Whether this reflects a real difference in the use of pottery in the later phase or just differences in the types of contexts excavated for each phase will only be answered by further excavation; but fine wares are far from absent in the pit-fill, and the painted designs are often more elaborate and more carefully executed in Phase II, even on the larger vessels. Close lattice motifs are common (e.g.Figure 26.5). Of particular note are two pieces with painted human figures, one on the interior of the base of a small vessel (2959.D1), the other on the outside wall of a pot of uncertain form (from 2910). In both cases the body is formed of two triangles joined point to point, giving the impression of a dress or tunic with a narrow waist. Almost certainly Mellaart's reconstruction of a human figure in his Fig.6: 13 should show this form of body rather than a single triangle. 

The most significant find came not from the pit-fill but from a recent collapse deposit in the base of the trench (2913.D1). However, given their position, these sherds are most likely to have weathered out from the pit, especially as they showed slight traces of the organic deposits found on other sherds from the feature. On the other hand, few sherd joins have been noticed so far within the pit-fill deposits, while these pieces form a large part of an open bowl, painted rather crudely on both surfaces. What sets the vessel apart, however, is the presence on the inside of the base of a scene apparently showing women dancing in front of a pair of larger figures (Figure 29). While single figures of the type mentioned above have also been found at Can Hasan (French 1962, figs. 5.12, 7.3), such a complex scene appears to be unique within the Chalcolithic of the Konya Plain. It certainly appears to be of local manufacture, with a typical fabric, slip and paint colours. Perhaps this vessel, resembling the figurative scenes from the walls of the Neolithic houses, emphasises the changing locus of decoration from domestic achitecture, usually seen only by the resident household, to portable objects that could be moved or exchanged between households or even communities. It is possible that such a shift is related to the development of more organised settlement systems and exchange networks in the Chalcolithic. 

These collapse deposits, like 2902 in Trench I, produced large quantities of pottery which had either eroded in from the original trench sections, or derived from the still in situ deposits cut back to create the step sections in each trench. Although lacking in specific contextual information, they do allow a quantitative comparison between the two Trenches. 2902/2915/2920 in Trench I produced over 100kg of pottery, while 2903/2909/2913/2916 in Trench II yielded almost 120kg. In total this represents over 50% of all the pottery found this year, and should not be ignored. They form the basis for our quantitative reanalysis of Mellaart's ceramic typology. 

Trench III

The remainder of the Trench III fill above 2947/2948/2970 was divided into four 0.2m spits (2924, 2926, 2930, 2933), from each of which five zembil of soil (c 150L) were dry-sieved (these samples were given separate Unit numbers: 2925, 2927, 2931, 2934). The sieving was designed to allow direct comparisons of densities and sizes of artefacts at different depths, as an assessment of the degree of post-depositional alteration of the deposits. The greatest quantity of pottery, and the largest sherds on average, came - rather unexpectedly - from the uppermost spit (modern ploughsoil); this may reflect the presence of more Classical material in the topsoil; certainly the amount of building material, which largely reflects Classical activity, is significantly greater in 2925 and 2927 than in the lower spits. Animal bone, which is more fragile, may be a better indicator of the intensity of post-depositional processes: the quantity of bone increases with depth, particularly between 2927 and 2931. Obsidian is also more common in the two lower spits, while the largest pieces of stone came from 2931. There is therefore a clear difference between the densities of different materials in the upper two spits (the modern ploughsoil and just below it) and the lower two, though all levels include small quantities of post-Chalcolithic material (e.g. a fragment of glass from 2934). This may reflect two different sets of processes altering the upper deposits of the mound during and since the Classical period. The best explanation might be episodes of colluvial slopewash, perhaps as a result of agriculture, combined with continued ploughing of the upper part of the deposit. 

This colluvium was found in the 1994 scrape squares on the eastern side of the mound but not in our trenches on the summit and in the south-west. It therefore seems to be restricted to the south and east of the site - though the northern and western flanks of the mound remain uninvestigated. However, a distinction between formation processes to the east and the west might be related to the differences in the density of Classical material discovered during the 1994 surface survey of the West Mound, which was significantly greater on the eastern side of the site. 

Trench IV

Conclusions and Comparisons

A full range of artefacts, faunal and botanical evidence was recovered which will allow us to compare Neolithic and Chalcolithic activities and begin to understand how lifestyles changed over the long term at Çatalhöyük. Many results are still awaited, but preliminary analysis has yielded some interesting observations. Faunal remains from pit 2941 in Trench II include relatively high numbers of large equids, raising the question of possible horse domestication at an early date. However, sheep/goat are the most common species; their size distribution indicates a population of mainly small animals with a minority of very large ones, suggesting a similar pattern of exploitation to that seen in the later levels on the East Mound. 

The obsidian industry likewise suggests continuity from the later Neolithic, with a number of good blades and some blade core fragments, perhaps contradicting Mellaart's description of the industry as 'poor'. Formal tools were not common, but include a transverse arrowhead of Levantine type, from Trench I. 

The worked stone shows an increase in both quantity and variety compared to the Neolithic; as well as the grindstones from the house floor, which have been sampled for residue analysis, finds include a very fine small polished stone bowl. 

Botanical evidence so far suggests a similar picture to the Neolithic, with cereals (wheat and barley), hackberries and reeds all present. Phytolith analysis of greenish lenses in the Trench II pit-fill has identified these as dung. Further botanical and chemical analyses are awaited, likewise information from micromorphology about activities on the floor of the Trench I building and the formation processes of the Trench II pit-fill. 

The vast majority of the finds from the West Mound comprised ceramics, mainly potsherds but also a number of potstands. Some of the pottery has been discussed above in relation to the character and phasing of individual deposits. These include some unusual finds from the lower part of the building fill in Trench I, differences between the extra-mural deposits 2923 (later, stratified) and 2939 (earlier, massive), and the suggestion that the Trench II building is typologically later than that in Trench I. In general, however, so much pottery was collected that most has not yet been analysed beyond basic quantifications. The new assemblages add to Mellaart's typology in a number of ways, having produced new forms and decorative motifs for both phases, and allowing quantitative analyses of vessel shapes, sizes and surface treatments, as well as information from sherd size distributions on the formation processes of different deposits. Of most interest in terms of media and publicity shots were the representations of human figures, particularly those mentioned from the Phase II deposits in Trench II but also one piece from the fill of Trench I (2902) which has moulded features in the shape of a face and recalls similar vessels from Hacilar. 

Further information on fabrics, little discussed by Mellaart, will allow the comparison of ceramic technologies between different phases, and with other sites. It was immediately clear, for instance, that the large incised potstands have a very different composition from the pottery vessels, being mainly coarse, porous, chaff-tempered fabrics. On the other hand, the horned 'potstands' from deposit 2952 are all in dense, mineral-gritted fabrics more akin to the pots. This may reflect a chronological difference: all the stratified examples of the incised type came from Trench II, mainly from pit 2941 but also from the (late) Phase I fill 2963. However, Trench I infill 2902 also produced a couple of fragments. 

For more detailed fabric information, a sample of body sherds has been exported for thin section analysis, which will allow comparison with the Neolithic pastes and perhaps provide some data on provenance. The regional and trans-regional context of the West Mound and its pottery is of some importance; as Özdoan (1993, 180) has pointed out, there are major problems in terms of our lack of knowledge of Chalcolithic sites in western and north-western Anatolia, and the 'bewildering datings' of pre-Bronze Age layers in Central Anatolia. Absolute dates for the West Mound are still lacking, with the exception of one from a geomorphologic core through the mound; the determination of 6940+90 BP broadly corresponds with the old dates from Can Hasan 2B of around 7000 BP. However, Late Neolithic/Early Chalcolithic material from a buried landscape at the base of the East Mound were associated with a date some 400 years older, which would close the gap between the end of the Neolithic and beginning of the Early Chalcolithic. The issue of whether or not there was a hiatus between the occupations of the two mounds will remain open until basal deposits under the West Mound can be sampled more extensively. However, a better synchronisation of the Early Chalcolithic of Çatalhöyük with Can Hasan and other dated sites could be produced by a programme of dating the deposits excavated this season, and it is hoped that resources will be forthcoming to provide this. Several radiocarbon dates for the East Mound were determined from seeds, and similar material is available from the West. 

In the mean time there is potential for topological dating. Comparative study of a sample of the painted pottery from Can Hasan has shown that fabrics there are generally very different: the majority of sherds are vegetable-tempered and surfaces are matt. At Çatalhöyük vegetable matter is a rare inclusion in the pottery (according to Mellaart [1965, 136] there is 'sometimes a little straw' in the coarser vessels of Early Chalcolithic I, but none at all [ibid, 156] in Early Chalcolithic II), while surfaces were usually burnished after painting - producing a blurred or washy appearance to the lines of paint in some cases. Hence the similarities seems to occur at the level of designs rather than technology. While French (1967) sees the floruit of the Çatalhöyük style at Can Hasan as restricted to layer 2B phase 2, layer 3 and the whole of 2B include shapes and designs reminiscent of, respectively Early Chalcolithic I and II, with the early phases of 2A showing some similarities with Early Chalcolithic II designs (e.g. striped jar necks), before a polychrome style, not seen at Çatalhöyük, develops in the later phases of 2A. These levels are equated with Mersin XXIV-XIX in the late 8th and first half of the 7th millennium BP (French 1966); a late phase of layer 2B also produced a tiny sherd of Halaf type (French 1968). 

A few sherds with incised and impressed decoration appear to be imports to Çatalhöyük, and have some resemblance to similar wares at Can Hasan - where incised pottery is relatively common, especially at the end of layer 2B - and elsewhere in Anatolia. In general terms they also resemble Balkan wares and therefore have relevance to an ongoing debate about relations between Anatolia and Europe during the Chalcolithic (Balkan Neolithic). In fact the Konya Plain, with elements and influences that can be termed both 'Halaf' and 'Balkan', appears to mark an area of interface between a Near Eastern cultural zone and an Anatolian-Balkan zone (Özdogan 1993). Of course these are nebulous concepts based on ceramic styles without references to people or social processes, but the Early Chalcolithic represents a period of expansion on the Konya Plain, if the number of sites identified during survey is anything to go by, and pottery decoration may have been one means of expressing affiliations and mediating contacts. Indeed it is hard to explain the sudden developments in all aspects of the ceramic repertoire compared to the Neolithic (decoration, the variety of forms and sizes, the variety of fabrics, absolute quantity of pottery) without seeing it in the light of developments in other areas. Technological change cannot be understood in a social vacuum. Hence, issues of continuity and change in other practices - architectural, economic, ritual - all of which the West Mound project has shed some light upon, are bound up with any discussion of ceramic typology. 

The unusual pieces fall into two groups (A and B). Group A consists of a few sherds with burnished surfaces and grooved decoration of several parallel lines in curvilinear patterns; group B has zig-zag bands defined by incised lines and filled with impressed dots or punctates. In both groups there is a little evidence for white incrustation of the design elements. Sherds of group A were found in Early Chalcolithic I deposits in Trenches I (2960 -Figure 30.3) and II (2963) as well as in the second spit excavated in Trench III (2926) and one of the 1994 scrape squares (856). In these last two contexts their associations were entirely with material of Early Chalcolithic I. The pieces of group B was differently distributed, deriving from recent collapse/fill layers in Trench II (2903 - Figure 30.2, 2921 - Figure 30.1), and a Classical pit in Trench I (2919), as well as from a surface context on the East Mound in 1997 (2401). The material from the Trench II contexts is overwhelmingly Early Chalcolithic II, that from 2919 Early Chalcolithic I while the other pottery from 2401 appears to be Late Neolithic in date with red slips and S-profiled forms but no painted motifs. Does this mean Group B material had a longer currency? Certainly they vary in fabric and in surface colours, the pieces from 2903 and 2401 having a moderate/high density of vegetable inclusions, and those from 2903 being dark grey while the others are buff or brown. Group A sherds are all dark brownish, some with a little vegetable matter. The stratified pieces from 2960 and 2963 both have roughly wiped or scraped interiors, quite distinct from the fine striations on interior surfaces of the painted pottery. Both groups therefore show a distinctive technology and decoration, which combined with their scarcity suggests they are indeed imports. 

The closest parallels for both groups is the site of Güzelyurt-Gelveri in Aksaray district (Esin 1993). It lies between Aksaray and Nigde, 150km north-east of Çatalhöyük, and close to Hasan Dag, long considered the source of the obsidian used in the Konya Plain. Limited excavations were conducted in 1990, during which pottery of five ware types were found. The most common (Ware 1) occurred in plain burnished (1a) and decorated (1b) groups. The decoration includes the parallel curvilinear grooves of Group A and the punctate-filled bands of Group B. Fabrics are predominantly chaff-tempered, and a few of the interior surfaces are scraped. While some of the 'Group B' material at Gelveri has broadly spaced, rather irregular dots, a large number of the sherds there have bands filled with regular rows of densely spaced fine impressions. These have not been found at Çatalhöyük. Moreover, the Group A sherds generally show regular design elements like the sherds from 2960 and one of the 856 groups, rather than the meandering lines of the pieces from 2926 and 2963. Hence, the Çatalhöyük sherds may not actually come from Gelveri, but they certainly belong to the same ceramic group. Similar material has been found to the north of Aksaray at the site of Kabakulak, surveyed by Ian Todd in the 1960's (Summers 1991). 

This Cappadocia connection is much stronger than any with Can Hasan, where the majority of the incised ware consists of broadly spaced lines in rectilinear designs (e.g. French 1962, fig. 5.13) or bands filled with elongated impressions rather than punctates (ibid., fig. 9.2). Similar decoration occurs on a sherd from the Büyükkaya at Bogazköy (Parzinger 1993, fig. 8.1). Chalcolithic material from north-central Anatolia has more affinities with Can Hasan than Çatalhöyük (Schoop, pers. comm.); the West Mound shows, through these few pieces, links to the eastern edge of the Konya Plain. It also implies that the Gelveri material is synchronised with Early Chalcolithic I. Esin's (1993) survey of similar material covers the whole period from proto-Chalcolithic Mersin to the earliest EB at Tepecik and Gelinciktepe; certainly sherds of Group B are seen at Mersin in level XXIV (Garstang 1953, fig. 36: 49-50). In the other decoration, Makkay (1993) suggests a link to the Pre-Cucuteni cultures of Romania (which also has vessels with painted human figures not wholly unlike those from Çatalhöyük - see Mantu 1993, fig. 4). However, this culture is dated in the first quarter of the 6th millennium BP, about 1000 years later than Can Hasan 2B. The significance of the Çatalhöyük finds may be to push the dating of Gelveri considerably earlier than its supposed Balkan parallels. 

Continued excavations on the West Mound at Çatalhöyük would be useful in order to provide further stratified material which may shed light on the suggested regional connections. The building partly exposed in Trench I is well enough preserved to repay further investigation; perhaps the 'potstand' deposit indicates this was a special area. Moreover, we still know almost nothing about the wider organisation of settlement on the West Mound, except that the buildings appear to be less densely packed than in the Neolithic. While burial practices also remain unknown, the presence of painted plaster in the building fill raises questions about the possible continuity of certain Neolithic practices. It is hoped that in future seasons more light will be shed on this intriguing site. 

Acknowledgements

Notes

References

French, D. 1962 Excavations at Can Hasan: first preliminary report 1961. Anatolian Studies 12: 27-40. 

French, D. 1965 Excavations at Can Hasan: fourth preliminary report 1964. Anatolian Studies 15: 87-94. 

French, D. 1966 Excavations at Can Hasan: fifth preliminary report 1965. Anatolian Studies 16: 113-123. 

French, D. 1967 Excavations at Can Hasan: sixth preliminary report 1966. Anatolian Studies 17: 1165-178. 

French, D. 1968 Excavations at Can Hasan: seventh preliminary report 1967. Anatolian Studies 18: 45-53. 

French, D. 1998 Canhasan Sites I: Stratigraphy and Structures. London: BIAA Monograph 23. 

Garstang, J. 1953. Mersin. 

Gibson, C. 1997 Konya Plain Survey Project 1996: report on the Neolithic and Early Chalcolithic pottery assemblages. Unpublished manuscript. 

Last, J. 1997 Surface pottery at Çatalhöyük. In I. Hodder (ed.), Çatalhöyük Excavations. Volume 1: On the surface. Cambridge/Ankara. 

Makkay, J. 1993 Pottery links between Late Neolithic cultures of the north-west Pontic and Anatolia, and the origins of the Hittites. Anatolica 19: 117-128. 

Mantu, C-M. 1993 Anthropomorphic representations from the Precucuteni and Cucuteni cultures. Anatolica 19: 129-141. 

Matthews, R. 1994 Archive Report. 

Mellaart, J. 1965. Çatal Hüyük West. Anatolian Studies 15: 135-56. 

Özdogan, M. 1993. Vinca and Anatolia: a new look at a very old problem. Anatolica 19: 173-93. 

Parzinger, H. 1993. Zur Zeitstellung der Büyükkaya-Ware: Bemerkungen zur vorbronzezeitlichen Kulturfolge Zentralanatoliens. Anatolica 19: 211-229. 

Summers, G. 1991 Chalcolithic pottery from Kabakulak (Nigde) collected by Ian Todd. Anatolian Studies 41: 125-131. 

Yakar, J. 1991 Prehistoric Anatolia: the Neolithic transformation and the early Chalcolithic period.
 

Figure 25 Plan of Trench I, and elevation of walls F300, F301.

Figure 26 Painted pottery from the West Mound (1-2 from 2960; 3 from 2958; 4-5 from 2910; 6 from 2964).

Figure 27 Horned 'potstands' from 2952.

Figure 28 Plan of Trench II (eastern section at top).

Figure 29 'Dancing Women' pot 2913.D1. 

Figure 30 Incised pottery from the West Mound (1 from 2921; 2 from 2903; 3 from 2960). 

© Çatalhöyük Research Project and individual authors, 1998