Zhusahn: A Small Neolithic Site Near Kokshetau, Kazakhstan
B. Bradley, S. Olsen, K. Akishev, S. Bektasov
In June and July of 2001, a small Neolithic site was investigated by the authors as part of the Krasnyi Yar Project, partly funded by the National Science Foundation (Grant # ). An excavation permit (open list) was obtained by Akishev prior to the initiation of excavation. This small site was discovered during a systematic survey of a transect between the Eneolithic sites of Krasnyi Yar and Vasilkovka, conducted by Eva Hulse. The site is located northeast of Kokshetau, Kazakhstan (Figure 1) about .5 km northeast of the Eneolithic site of Krasnyi Yar. The site is on the margins of a relict pond/forest and extends up the low sloping side of a shallow draw (Figure 2) in an open area in the forest/steppe. Although the majority of artifacts at the site were recovered from surface contexts, we obtained a good sample of Neolithic artifacts to compare to the assemblage of the nearby Eneolithic site of Krasnyi Yar. We investigated several research questions including: 1) Does the site location in relation to land forms reflect what Neolithic people were doing there?; 2) Are there subsurface features present?; 3) Are all of the artifacts the result of Neolithic use, or are there other time periods represented?; and 4) How does the artifact assemblage, technology, and raw material compare to those from the Eneolithic site of Krasnyi Yar?
Zhusahn was initially discovered during a survey project directed by Eva Hulse. An initial pass across the site area found only a couple of isolated artifacts, and the area was not recorded as a site. Additional inspection by Hulse revealed a wide scatter of artifacts including flaked stone, ground stone, and pottery. At this time the site was designated and recorded. Because of the presence of sage in the site area, it was named Zhusahn, the Kazakh term for sage.
Initially, artifacts were plotted using a Global Positioning System, but not collected (Figure 3) . Once we decided to do more intensive investigation at the site, the artifacts located during survey, plus many others, were relocated in relation to a grid system with a total station.
Mapping and Excavations
After mapping surface artifacts, two concentrations became apparent, although these concentrations were not dense. Artifacts were collected and taken to the laboratory where Bradley did initial analysis and comparisons to assemblages in the local archaeological literature. Although there were clearly different time periods represented, based mainly on pottery, it was determined that the majority of materials were probably of Neolithic origin. This was especially true of the two artifact concentrations. We decided that further investigation would be worthwhile, especially in establishing a baseline comparison between local Neolithic and the nearby Eneolithic Krasnyi Yar site.
The presence of a Neolithic site on the margin of what was once probably a small forest-surrounded pond was also intriguing. This is not a usual Neolithic site location in the area. Typically, Neolithic sites are found along rivers and on lake margins. We decided to undertake a small excavation to increase our artifact sample and to determine if there was any potential for subsurface site features.
We selected one of the two surface artifact concentrations for excavation. This was the concentration closest to the relict forest-pond and lowest on the hill slope. We figured that this area had the greatest potential to yield in-place, undisturbed deposits. A 5 X 3 meter area was laid out oriented to the cardinal directions (Figure 4) . Excavation was done by meter squares, using arbitrary levels. Excavations were halted once sterile, non-cultural deposits were encountered. Two 25 cm balks were left through the excavation unit, one oriented north-south and the other east-west, to retain evidence of stratigraphy. Individual artifacts were mapped in-place as they were found, and all deposits were passed through 6 mm mesh screen to recover dislodged and small pieces. Individually mapped pieces were collected and bagged separately with their coordinates, while dislodged and pieces found in the screen were kept together by grid square and level.
After most of the grid unit was excavated, it was determined that all artifacts were being recovered from recently disturbed deposits, specifically a plow zone. We decided that we needed to examine the stratigraphy both up the hill to the northeast and downhill to the southwest. We hoped to determine whether or not the artifacts we were finding may have washed in from up hill, and if there might be undisturbed buried cultural deposits down hill. A trench 1 meter wide was excavated down to a marl-like deposit in both directions. The stratigraphy in both sections mirrored that in the gridded area. There is no evidence that artifacts had come from uphill or that there are buried cultural deposits downhill.
Upon completion of excavations, all areas were filled with a mixture of sediment that had been removed. No attempt was made to return fill to its original location or to reconstruct the stratigraphy. This mixture of sediments should make our excavation boundaries evident in the case of future work at the site.
Sediment deposits in the areas of our excavations are typical of a steppe physiography. The small pond/forest is located in the bottom of a shallow draw, draining to the northwest. Most of the artifacts were found on the pod/forest margin and the lower areas of the southwest facing side of the draw. Throughout our excavations there was ample evidence of relatively recent plowing down to about 50cm in depth. This plowing not only mixed the upper humic A-horizon soil with underlying sediments, it formed soil cracks that extended down another 50cm in places. These soil cracks did not occur in the southwestrn area of our excavations where it hadn=t been plowed (Figure 5) , (Figure 6) , and (Figure 7) .
Bedrock in the area consisted of a marl-like material that varied in color from white through ochre yellow to red-pink. In one area, this material was full of small mica specks. The hill side to the northeast changed gradually from a silty loam deposit to one full of small colluvial pieces of granite. These were size-graded from larger toward the top of the slope to small toward the bottom, dropping out altogether before the edge of the pond/forest. Artifacts were found throughout the plow zone and down into the soil cracks below. It is probable that all of the artifacts were on the modern ground surface at the time it was plowed.
The majority of artifacts recovered from Zhusahn are flaked stone. Considering the disturbed nature and shallowness of the deposits, it is not surprising that this is the case. Only a few scraps of animal bone were found and few of these are identifiable to species. One exception is a horse tooth that was recovered from the plow zone in the stratigraphic trench. We also recovered a number of pottery sherds, but most were eroded due to long surface exposure and chemical weathering. In addition, we found a few ground stone artifacts and one that was flaked and then ground. Since there are different time periods represented by the pottery, it is likely that this is also true of the stone artifacts. However, with the exception of only a couple of pieces, nothing we found is considered distinctively diagnostic of only a single time period. Because of this, we have chosen to analyze and describe the flaked stone artifacts as a single assemblage. Even if there is some mixing, it should be clear what the underlying technologies were and what was taking place at the site. Items thought to have originated from other than the Neolithic use of the site are noted.
Several different stone sources were used as flaking material at Zhusahn. The types we recorded are agate, fine chert, coarse chert, jasper, fine quartzite, coarse quartzite, siltstone and other (see Table 1) . Only the siltstone source area has been identified to date. Although we recorded several types, it is likely that two of them are actually the same source, but of different quality. Specifically, this is the fine chert and fine quartzite. Similar ranges of variation, from a fine grained chert through a fine quartzite, are common in North American materials, such as Beaver River sandstone from northern Alberta, Canada. The differences are related to the degree of secondary silicification. It is also apparent to us that the coarse chert and coarse quartzite are different sources from each other as well as from the fine chert and quartzite. Therefore, at Zhusahn, there seem to be five different stone sources present.
A distinctive siltstone has been recovered both at Zhusahn and at Krasnyi Yar. We have also observed it on other sites in the Kokshetau region. This material comes from the top of a hill, on the west side of Kokshetau, not far from the village of Sadovaya. Although recent quarrying activity has impacted some of the source, much of it remains. We recorded this source as a site, named Zhartas after the Kazakh word for stony hill. Of particular interest is a series of 31 small prehistoric quarry pits arranged around the top of the hill, and a concentration of flaked stone artifacts. We saw evidence of both biface and pressure blade manufacturing. According to a surface geology map, this material is of Cambrian origin. It ranges from well-silicified high quality siltstone to an unsilicified mudstone. At least three distinct colors are present, gray, green, and dark red, and they come from different areas of the quarry site. This quarry is within 10 kilometers of both Zhusahn and Krasnyi Yar.
Flaked stone artifacts that show no evidence of retouch for tool use are known as debitage. Within this category are a number of artifact forms that may be determined to reflect their manufacture technology. From this information it is possible to infer what kinds of flaking were occurring at Zhusahn.
Angular debris are flaked stone fragments that do not retain any direct evidence of their technological origin. These tend to occur during many types of flaking and are a byproduct, usually unintentional. Large quantities of angular debris on a site may indicate that a significant amount of flaking was occurring there. Only 21 pieces (5% of debitage) of angular debris were recovered at Zhusahn. This indicates that core reduction/flake production was not commonly practiced at the site. Coarse chert, coarse quartzite and agate are absent in the angular debris.
Biface flakes are produced during the bifacial reduction of a piece, whether during the production of an axe, an adze, a projectile point, or other tool. They are distinctive in form. This is not to say that all flakes that come from a biface will have this form. Unless biface flakes are used for tool blanks they were mostly waste material and generally indicate that biface manufacturing or at least reworking was taking place at a site. Zhusahn yielded 22 (6%) biface flakes (Figure 8 f-g) . Their size indicates that most probably resulted from percussion flaking. These flakes are mainly fine grained materials, but surprisingly none are siltstone.
We distinguished two other types of flakes that are usually the result of generalized flake production from cores, or from the early stages of biface manufacture. These are core flakes and disc core flakes. The former come in many shapes but tend to be either slightly elongated of have nearly equal length and width dimensions (Figure 8 a-e) . Disc core flakes are usually wider than long and have thick platforms. They have a distinctly triangular section. When either type makes up a significant proportion of an assemblage, it usually indicates that generalized flake production was present. Curiously, although core flakes represented 38% (151) and disc core flakes 6% (23) of the debitage, we only recovered a single flake core and it was simply a tested piece of siltstone. The relatively low count of angular debris plus the lack of flake cores probably means that most of the flakes were coming to the site already made. One may argue from this that many were used as tools, and we guess that traceological analysis would bear this out.
Zhusahn also yielded 151 (38%) blade fragments (Figure 9 c-bb) . Blades are specially made flakes that are at least twice as long as they are wide. In the case of Zhusahn, the blades were produced from specially shaped precores, were made by pressure, and the final cores were bullet-shaped. We distinguish several types of blades including corner blades, side blades and center blades. These types inform us about the specific blade technology being used as well as about what forms of blade cores were being imported. The process of making regular center blades produces proportionally larger numbers of corner and side blades, especially early in the process. If all of the blade core manufacturing was taking place at the site, we would expect corner blades and side blades to significantly outnumber center blades. Early in the process, corner blades are common as the core is formed. Few center blades are produced. As the core becomes regular, fewer corner blades are produced, until none are made as the core becomes round. Side blades occur mostly during middle stage production but continue to be made right to the end. Even when a core is well-made and the blade production sequence proceeds without major errors, side blades will equal or exceed center blades. These general observations hold true for any blade making technique, but the control seen in pressure blade production tends to moderate these tendencies. Never the less, Zhusahn debitage includes only 6 (9%) corner blades and 18 (25%) side blades. The remaining 47 (66%) are center blades. The inclusion of tools made on blades does not change the relative proportions. Two blade fragments (Figure 9 a-b) are large and relatively thick. Both were made by percussion and have weathered surfaces. One has opposed notching, possibly from use as a wedge. Technologically, these are unlike the pressure blades and they may be of Upper Paleolithic origin. If this is the case, there may be an Upper Paleolithic site somewhere in the area.
There are only three other types of debitage; core tablet, core, and other. Core tablets are flakes struck from the top of a core, usually a blade core, to rejuvenate the flaking platform. The presence of core tablets at a site probably indicates the production of blades there, unless the core tablets have been modified into tools. All three core tablets (Figure 8 h-j) at Zhusahn were unmodified and look to have been struck from cores well along in their use life.
Two pressure blade cores (Figure 8 k-l) , one complete and one fragment were found at Zhusahn. Both were used up. The complete one has a remnant of cortex on one side indicating that it never achieve a bullet shape. The other is a distal fragment that clearly indicates it was very regular and bullet-shaped. Both were made from fine chert.
Evident in the various forms of debitage are characteristics of intentional heat treatment. In some cases, the controlled application of low heat to flaking stone will change its structure, making it more glassy and easier to flake. Care has to be taken not to overheat it, which will destroy it. Heat treated stone at Zhusahn shows a glossy texture and/or a change in color. The lighter cream colored chert takes on a pink tinge, and in some cases turns red. Heat treatment was being done to both bifaces and pressure blade cores.
In conclusion, the flaked stone debitage from Zhusahn indicates that three forms of stone were being brought to the site; ready-made flakes and pressure blades, and pressure blade cores that were already well along in their use. Undoubtedly some tools were being flaked on site, especially large biface tools, and at least a few blades were also being produced. It seems, however, that the bulk of flaked stone was being brought to the site already finished.
Stone raw materials are dominated by fine cherts and quartzites of unknown origin, but jasper, agate, and coarse cherts and quartzites are present, also of unknown sources. It is curious that the nearby and readily available siltstone is only meagerly represented, even though there is evidence at the source of exploitation during the Neolithic. In order to understand what was going on at Zhusahn, it is imperative that the other sources of stone be located.
Twenty-four stone tools were recovered from the surface and excavations. Two of these are ground/polished and the rest are flaked (one ground after flaking). They are classified into general classes, mainly based on their inferred use(s). Only one of the tools, a whetstone, is clearly from post-Neolithic times, but some of the others may be as well.
Three projectile points, two fragments (Figure 10 e-f) and one in two pieces, were found at Zhusahn. The two fragmentary points are distal ends and both exhibit evidence of impact breaks. Both impact breaks are the type that may be expected from high energy contact with large dense animal bone. It seems that they were used for hunting large animals. Although bifacial projectile points are found in Eneolithic sites, such as nearby Krasnyi Yar, both of the pieces from Zhusahn exhibit different manufacture technology. One was made from a fine grained tan-brown chert and the other from jasper. Both have flat cross-sections and exhibit controlled, shallow, well-spaced pressure flaking, followed by abrupt and very fine, non-invasive pressure retouch. Both points were extremely thin and well-made. Although only fragments, these pieces seem too large to have been arrow points. They may have been used on thrown spears, either by hand or with a throwing stick.
The third projectile point was found in two pieces that refit (Figure 11e) . It was made on a large blade or blade/flake and exhibits pressure flaking on both faces. There are original flake surfaces left on both faces. This pieces was broken during manufacture, possibly on a fault left over from the flake blank. Pressure flaking on the dorsal face is selective and some of the flake scars are diagonal (a technique typically used to remove prominent blade scar ridges). Pressure flaking on the ventral surface is flat, perpendicular to diagonal, and at least one flake removal was an over-shot. The size and form of this point preform is reminiscent of some of the pieces from Krasnyi Yar, but we are unwilling to call it Eneolithic. This piece was probably being made at the site because it is a manufacture break and we recovered both pieces within a single meter square.
A very small fragment of a fourth possible projectile point was recovered as well. It is such a small piece that all that we can say about it is that it was made from red chert (heat treated?) And pressure flaked on one face and percussion flaked on the other. It is possibly a fragment of a flaked knife rather than a projectile point.
Only two small bifaces were found (Figure 10 a-b) and it is unclear what they were intended to be. One was made of red chert and the other from a gray-green medium-grained quartzite. Both exhibit only percussion flake scars. Neither is large enough to have been made into projectile points.
Two bifacially flaked drills (Figure 10 c-d) , one complete and one fragment were found on the surface at Zhusahn. Both are relatively thin and neither shows obvious evidence of rotary use. They are identified as drills based entirely on their forms. Both were made of gray chert. Thier forms suggest that they may have been hafted for use.
The most common tool type is small scrapers. We recovered 20 scrapers made on several blank types (Figure 12) . Three were on blades (2 on corner blades), 14 were on core flakes, one was made from a possible core tablet, and one was formed on a natural small tabular piece of stone. Materials were typical for the site with 12 scrapers made on fine chert, 1 on coarse chert, 4 from jasper, one from fine quartzite, and one from siltstone from Zhartas. Most have single retouched edges that are sharp but abruptly retouched. Three look like they may have been broken during use (in a haft?), and one is heavily use-damaged. Most of the others retain sharp useable working edges. Overall these are fairly small scrapers with a range in length from 17 to 46 mm. Their mean length is 24 mm. There are no distinctive forms and any of them could be found in Eneolithic or Neolithic sites.
Two large end scraper-like tools and a fragment were found (Figure 13 d-e) and (Figure 14 b-d) . Both exhibit percussion retouch all around their edges. Although they may have been used as scrapers, their size and forms may indicate that their intended use was as hafted adzes. This interpretation may be supported by the fact that both were made from tough durable chert. Finer grained, more brittle material, would be less suited to adze use.
A single flaked stone object (Figure 13b) and (Figure 14e) shows rough percussion flaking to form an irregular tool with a plano-convex cross-section. Its irregularity my indicate that it was unfinished. The form looks like its intended use may have been as an adze or chopper.
A single burin (Figure 11a) was made on the midsection of a relatively large blade. It has a single burin flake scar. It is possible that this was not intended to be a formal burin. It was made from a fine-grained brown chert.
Two artifacts have been classified as axes. One may be either a rough flaked axe or possibly an axe blank (Figure 13a) . It is also possible that this specimen was some type of chopper or possibly used like a hoe. It was shaped from a gray-tan coarse quartzite. The second piece is just a bit fragment of a beautifully ground and polished axe (Figure 13g) . It was made of banded gray slate and was ground to a sharp even edge. The break is the type that occurs during impact with a strong but yielding substance, possibly wood. This fragment was recovered from the southwest end of the stratigraphic trench (Arb Unit 2), at the edge where we think there was a small forest.
The base or tip of a flaked then ground artifact resembles a fragment of a pick (Figure 13c) and (Figure 14a) . It could also be the base of an axe or even a hoe. It was made from a gray-green silicified shale.
A curious pebble tool was found on the west edge of the surface scatter of artifacts. It is a flattened, rounded rectangle (Figure 14f) . All faces exhibit scratching in multiple directions. The stone type is unknown but looks like it may be a coarse limestone. It is finely abrasive. At one end there are partially drilled conical holes, probably made with a stone drill. The holes are aligned but do not meet. This piece is reminiscent of Iron Age sharpening stones or whetstones. This attribution is not unexpected because at least one of the fragments of pottery found in the same area is probably Iron Age.
A single piece esquilleee (splintered piece) came from the excavations (Figure 11b) . It is the same cream-colored fine grained chert that was commonly used for the pressure blades. This type of artifact is found in many time periods around the world and is thought to be the discarded remnant of a flake used as a wedge. Pounding on one edge with the other in contact with a hard surface, produces shattering. These artifacts are sometimes difficult to distinguish from bipolar cores.
Two flakes were found that have retouch along one side. The retouch is not formal enough for them to be classified as scrapers, but one is fragmentary (Figure 11d) and may be the proximal broken end of a formal scraper. Both were made from fine-grained chert.
Five artifacts are fragments of blades that have had one ot both edges modified by retouch (Figure 11c) . One has a notched edge but the notch may be accidental or the result of use. All of the pieces could have been inserted in handles or even composite projectile points. It is unclear if they were used singly or in tandem with similar pieces. It is also very likely that many of the unretouched blade fragments classified as debitage were also put to the same use. It is clear that the primary technology at the site was the production of pressure blades. Surely, this was for more than the retouching of just a few fragments. If these artifacts had been recovered from a good sealed context, it may have been worthwhile to have them analyzed for wear.
A single small piece of igneous stone may be a fragment of some type of ground stone tool. Unfortunately the piece is too small to make a determination.
It is likely that the tools at Zhusahn represent various activities, possibly at different times and even in different cultures. The two impact broken projectile point fragments may be Neolithic in age, but their presence and near proximity to each other, may indicate the location of a large animal kill. The bone are long gone. Typically, projectile point bases are found at campsites and tips and midsections are found either at the kill or the nearby butchering location.
Hide processing is also indicated by the small scrapers that were the most abundant formal tool. At least three of them were broken during use, probably in their hafts. These, plus the one that exhibited heavy battering wear, may be evidence of wood working, with the small scrapers hafted as adze blades.
The large number of blade fragments, some with retouch, may indicate that the site served as a butchering station, if the blade pieces were inserts in composite knives. This tool form is known from hafted specimens in Europe and various uses have been identified. There is no evidence of sheen on any of the blades from Zhusahn, so it is unlikely that they came from sickles. It is also clear that most of the flakes came to the site as flakes rather than being made there. This is consistent with a short-term multiple visit origin for the assemblage.
Zhusahn is a site with a widely scattered artifact assemblage, around the perimeter of an old pond/forest. All of the artifacts we recovered were probably originally on the surface, with plowing accounting for the subsurface finds. Technologically, the assemblage is mainly made up of pressure blades and the debitage resulting from their manufacture, most likely of Neolithic age. There is also a significant component of core flakes that were evidently brought to the site already made. Although there are some tool types shared with the later Eneolithic, such as the small scrapers, the pressure blade technology is absent from nearby Krasnyi Yar. This Neolithic assemblage is different enough from the Eneolithic that it suggests that they may not have been historically related.
The proximity to the pond/forest may indicate that the site served as a hunting and gathering station, rather than a camp. The lack of features or even fire-cracked rock also supports the idea that the site was a limited activity area. Most site use occurred during the Neolithic but whatever happened, the site location attracted people off and on over many hundreds, and possibly thousands of years.