Abstract: The structure of a lithic assemblage and the nature of artifact variability are closely related to several factors, including the availability and quality of raw material, the strategy by which it was procured, the particular activities in which stone tools were made and used, and the role of the sites within a settlement system.This paper presents the results of an analysis of raw material exploitation and their implications for hominid adaptations at Zhoukoudian Loc.15, an important late Middle and early Upper Pleistocene site in North China.
Raw Material Utilized at the Site
Six lithic raw material types are recognized at the site, namely vein quartz, igneous materials, rock crystal, flint, sandstone, and quartzite.The overwhelming majority(95.2% )of the artifacts were made on quartz.Among the minority raw material group, igneous rock is most abundant (204 pieces), followed by crystal (81 pieces).Only 30 pieces on flint, 16 pieces on sandstone, and 2 pieces on quartzite were collected.
Raw Material Distribution, Availability and Quality
Most of the lithic materials exploited at Loc.15 were available in the landscape close to the site. Longgushan and the larger Xishan where the Zhoukoudian complex is situated, are composed of various rock types, including quartzite, limestone, sandstone, and slates.The weathered outcrops at the hills and the riverbeds of the Zhoukou River provided a ready source for hominids to find and select the stone materials they needed for making their tools.A gravel layer in the Zhoukou River Terrace, the so-called “the Lower Gravel”, formed during the Middle Pleistocene, is rich in various kinds of rounded stones and is believed to be one of the principal raw material sources for Zhoukoudian hominids. Nodules of vein quartz, quartzite, igneous rock, and sandstone can be easily encountered in the area; rock crystals have been found in a granite area about 5 km north of the site.A few exotic materials, such as flint and agate, cannot be easily located nearby.However, such materials make up only a tiny proportion of the assemblage, and they were probably procured as small isolated nodules in streambeds around Zhoukoudian. While numerous stones can been easily obtained close to the site, materials with high quality are very scarce in the region.In general, workability of the dominant raw material types at the site, namely vein quartz and rock crystal, is very poor and thus limits the number of usable flakes detached from any one core.Quartz tends to split along structural planes and thus makes it difficult the toolmaker to control the size and shape of the flakes detached from it.
Strategies for Exploiting Different Raw Materials
Quartz is the major material used to produce cores, flakes, and tools.Both direct hammer per- cussion and bipolar flaking were applied to this kind of material.Only 2 hammer percussion cores (flint)were collected for the non-quartz materials; a much smaller proportion of debris was found from these material compared to quartz; and flakes on these non-quartz materials are larger and more regular.These might indicate that most flakes on flint, igneous rocks and sandstone were not produced at the site; they were detached at other spots and the selected materials were then brought to the site. This conclusion is also supported by the result of Minimal Nodule Analysis:non-quartz materials were sorted into minimal groups and refitting was attempted for them.No conjoins were found. Raw material frequencies were also examined for tools by class.More than 93.4% of the re- touched pieces were produced on quartz.This material was used to make all kinds of tools except for cleavers.More than 4% of the tools were produced on igneous material, including burins, chopper- chopping tools, cleavers, notches, and sidescrapers.Fourteen pieces were made on crystal, including burins, nail-shaped scrapers, and sidescrapers; 12 tools, all sidescrapers, were fabricated on flint. The 4 sandstone pieces are chopper-chopping tools.Only 1 retouched piece(a cleaver)was made on quartzite.Such statistics indicate that hominids at the site had the ability to select different materials to produce a variety of implements, but tended overwhelmingly to use quartz.
Consumption of Raw Material
Several measures have been used to assess the extent to which a piece of raw material or raw materials as a whole was consumed.
Core Reduction Extent
Core reduction extent or intensity was examined by the ratio of simple, minimally worked or tested cores to extensively worked cores.In the case of Loc.15 assemblage, the ratio of simple cores (23 pieces)to the combination of discoid (33 pieces)and polyhedral(74 pieces)cores have been used.A ratio of 1∶4.7 means that there are far more extensively consumed cores than minimally worked or test- ed cores.
Raw Material Use Intensity
A major measure of raw material use intensity is the ratio of unmodified flakes to retouched tools. A totally of 530 unretouched whole flakes and flake fragments were collected from Loc.15, and 848 tools retouched on flakes and flake fragments were identifid.The ratio of unmodified flakes-flake frag- ments to retouched flakes-fake fragments is 1∶1.6. There are two possible explanations for the high ratio of unretouched flakes to flake tools.One is that the large number of flake fragments and chunks at the site provided more usable tool blanks than the whole flakes.A total of 685 tools were modified on flake fragments, another 321 pieces were fabri- cated on chunks, compared to only 163 pieces made on whole flakes.The poor quality workability of quartz could explain the scarcity of whole flakes and the abundance of flake fragments and chunks. Some of such fragments might be thin and sharp, and thus were selected by hominids to make tools.A more plausible explanation is that, although many flakes were not retouched, pieces with natural sharp edges might have been used directly as tools, rather than being modified.“Utilized flakes” have been recognized in many Paleolithic sites, both in China and other parts of the Old World.However, it is always difficult to test if a piece of un-retouched flake has been used or not, especially with coarse- grained materials that prevent them from being analyzed for use-wear.Among the flakes from Loc.15, 66 pieces exhibit damage scars.Because those samples were not collected with intention of protecting the edges, use-wear examination on them will not be very meaningful.
The Index of Retouch Length
In order to assess the extent of retouch on the blank, an Index of Retouch Length was created for the modified tools.The IRL is simply the ratio of retouch length or edge length to the whole length of the margin at which retouch was located.The margin length was measured only for the part that has the condition or potential to be retouched into working edge.The larger the retouch index, the more com- plete the retouch, thus the more intensively a piece of raw material was modified.Only scrapers were included in this analysis.While some pieces were only minimally or partially retouched, an index mean of 0.89 indicates that for most of the scrapers, a large proportion (almost 90% )of the margin was retouched into workable edge.
The Index of Sharpening
An Index of Sharpening or resharpening was adopted in this study.The IS, calculated as the ratio of the edge thickness to piece thickness at the central ridge, would indicate how close the modifica- tion came to the central ridge and estimate the amount of material removed by primary retouch or re- sharpening from the original blank:the larger the index, the closer the retouch scars moved to the central ridge, and the more complete and extensive the modification carried out.Only scrapers made on flakes were included in this analysis.An index mean of 0.66 indicates that most of them were re- touched to a moderate extent.The majority of the pieces have the index between 0.4 and 0.8. A high IRL and a relatively low IS imply that while hominids at Loc.15 intended to make use of the lateral margin of a piece of blank as much as possible, they did not put much effort to rework their tool edges, a conclusion that is also supported by the fact that the overwhelming majority of the tools exhibit only one retouch layer.A possible explanation is that it was easier and more efficient to pick up a suitable piece of material, make use of its sharp edge or apply primary retouch to its lateral side to make a working edge, than to resharpen the worn edge on a used tool, provided that raw materials or tool-making potentials were abundant at the site.
The Number of Retouched Edges
When tool edge wore out, stone toolmakers might have several options to solve the problem.They could resharpen or rework the worn edge; create another edge or edges on the piece, or they could just abandon it and make a new tool.The selection of certain option(s)was closely related to raw material availability and the cost of getting new materials.The 1, 188 scrapers collected from Loc.15 preserve a total of 1, 313 retouched edgs.The large majority of the pieces (1, 063 pieces)are single-edged tools; The ratio of single-edged pieces to multi-edged ones is 8.5∶1.Such a high ratio means that in general, hominids at the site chose to replace the worn tools with new ones rather than reworking the used pieces. An analysis on the number of retouched edges on different materials reveals that these raw materials were not treated the same way.The ratio of single-edged pieces to multi-edged ones is 9.6∶ 1 for quartz, 3∶1 for crystal, 2.2∶1 for igneous material, and 1.4∶1 for flint, meaning that the non- quartz materials were more extensively consumed than vein quartz, an indication that hominids at the site would favor better quality raw materials when they were available.
Conclusions
From the perspective of raw material economy, the Loc.15 assemblage can be summarized as the follows: Raw material exploited at the site was locally available and was predominantly quartz, and is characterized by high abundance and low quality.The cost of obtaining raw material and using them to produce stone tools was balanced by these two conflicting factors, which in turn had strong effect on the nature of lithic technology at the site. The extent of raw material consumption in general is low, evidenced by the predominance of waste debitage or chunk fragments and the minimal modifications on the retouched pieces. Many flakes were not modified although some could have been utilized directly. Different strategies were involved in dealing with different raw materials.Quartz and rock crystal were processed at the site, and other materials may have been flaked somewhere else and selected materials brought into the cave.The amount of waste quartz and rock crystal is far higher than that for other materials.Quartz and crystal were less extensively consumed than other materials. Several ethnoarchaeological and experimental models developed in the West were applied to the study of raw material exploitation strategies and economy at Loc.15, such as Binford' s model of settlement organization dictating the production of curated or expedient tools, Kuhn' s notion of technological provisioning, Andrefsky' s relationship between raw material abundance-quality and formal-informal tools.The Loc.15 industry has found to fit the pattern of poor quality raw material leading to informal stone tools.The structure and an expedient feature of lithic assemblage at the site might also correlate with a more-or-less stable settlement pattern at the Zhoukoudian complex and the strategy of provisioning site possibly adopted by the Loc.15 hominids.

Key words: Raw material; Exploitation strategies; Paleolithic; Zhoukoudian Licality 15