2. Production and Distribution of Stone Spades

The full-coverage regional survey has helped us to reveal settlement patterns and identify important regional centres and craft production sites (Liu et al. 2002-2004). We have located four sites that appear to have specialised in the production of stone spades, based on the presence of abundant stone blanks. These are Zhaiwan, Huizui, Xiahousi and Xikouzi in Yanshi county, all located near the Songshan Mountains (Fig. 2). To understand in more detail the process of lithic production during the formative period of early states, we have investigated the locations of lithic resources, the raw material procurement methods, manufacturing processes, and distribution of finished products. We have particularly focused on the Huizui site using intensive survey and excavation, and the research has revealed great insight into the organisation of manufacturing processes for producing stone tools.

2.1 Raw material quarry

Various lithic raw materials were used for making different types of stone tools in the Yiluo basin. At Huizui, for example, spades were mainly made of dolomite; axes, adzes and chisels made of diabase; knives and sickles made of micaceous fine sandstone; and grinding slabs made of sandstone. These materials appear to have been locally available as river cobbles or outcrops in the mountains, which can be found in areas 0.5-10km from Huizui (Ford 2004; 2007; Webb et al. 2007). For making spades oolitic dolomite was the chief material, and sandstone was indispensable for grinding.

The raw materials of oolitic dolomite were quarried from dolomite outcrops along the northern ranges of the Songshan Mountains, about 4-5km to the south of lithic production sites (Figs 2, 3.1). The oolitic dolomite is thin-bedded, and breaks readily along parallel bedding cracks into thin slabs (Fig. 3.2). Therefore, such a raw material could have been easily quarried with Stone Age technology before the adoption of metal tools or explosives. Its thin-bedded formation on the outcrops provides the right thickness for fabricating into thin and long spades (Webb et al. 2007). Many outcrops appear to have been quarried, perhaps repeatedly, in the past, but it is difficult to date these activities. As these quarries are often located on steep slopes, any artefacts tht may have been associated with a quarry would have been easily eroded from their original locations or could have fallen into the gullies.

Numerous pieces of oolitic dolomite have been found at Huizui (Fig. 4.1). Examination of thin-sections of the oolitic dolomite using a petrological microscope reveals that it has a granular texture of interlocking dolomite crystals. The nature of the oolitic dolomite is one of the reasons it was selected to produce stone tools. Its homogeneous structure of interlocking crystals means that it will generally break evenly and predictably when it is being worked. Dolomite is a soft material (hardness 3.5-4) and is easily ground by harder rocks (e.g. quartz sandstone) to give sharp, effective edges and very smooth faces that can take a high polish. However, its softness means that it is readily scratched, and the thin edges of dolomite tools would chip if struck against harder rocks (Webb et al. 2007). Many spades uncovered from archaeological contexts are broken and their edges are badly damaged. Such drawbacks of this material are confirmed by an experimental study on digging with spades, in which one spade broke before it had been used 200 times (Owen 2006).

For making ground stone tools sandstone is essential. A large number of irregular sandstone fragments recovered from Huizui may have been used to abrade tool faces (Fig. 4.2). These are made of a variety of sandstone types: fine, medium and coarse-grained, and composed predominantly of quartz, although the coarser-grained ones generally have a substantial component of weathered feldspar grains (Webb et al. 2007). The coarser sandstones were probably used for initial grinding of the tool, with the finer sandstones employed to smooth surfaces and sharpen edges. Final polishing would have required a still finer medium, using very fine sediment (clay/silt), animal and/or plant products (Owen 2007).

The Lower Triassic sandstones outcropping in the two hills located near Huizui (Fig. 2) differ in colour and composition. The outcrop c. 1.5km to the south-west of Huizui is white, medium-coarse grained and feldspathic (Fig. 3.3), while the hill to the north-east, which lies within 500m of one of the excavation sites, is composed predominantly of fine-grained quartzose ferruginous violet-coloured sandstone. Both outcrops show evidence for quarrying, although much of it appears to be rather recent. Petrological examination under a microscope showed that the sandstone fragments found at Huizui encompass a variety of lithologies (feldspathic and quartzose, fine to coarse grained, often ferruginous), and were probably sourced from both hills (Webb et al. 2007).

2.2 Spade preforming

Our surveys have located a site that shows characteristics of the early stages of manufacture after raw material procurement. The site of Jiulongshuiku (about 1500 sq. m in area) is located on a foothill, about 3km south of Xikouzi, one of the spade production sites. It is also situated less than 100m north of an ancient oolitic dolomite quarry. The site is a relatively flat area, and the bedrock is composed of fine-grained dolomite, which is quite different from the oolitic dolomite used for spade manufacture, and does not show any evidence of quarrying. The lithic objects scattered on the ground surface are mostly non-local, including stone balls and hammerstones composed of sandstone, as well as some oolitic dolomite flakes (Fig. 5). This site may have been a locale where stonemasons flaked the oolitic dolomite preforms after quarrying.

This spatial pattern of extraction and reduction of raw materials finds its parallel in the lithic manufacturing process among several Aboriginal groups in the Central and Western Deserts of Australia. Small chipping stations at locations separated from the quarries were identified as locales for preforming cores and flakes, recognisable by the presence of hammerstones and other lithic debris. This stage of manufacture was employed in order to reduce the load for the return journey to the habitation site (Gould 1978, 819; 1980, 123). The lithic production system of these Australian Aborigines, which is placed at the lower end of the exchange continuum from production by consumers to commercial industrialisation (Torrence 1986, 51), may not be comparable to the Erlitou culture, but this quarry method is highly relevant to our case study. Given that the oolitic outcrops being quarried normally occur on the steep slopes of the mountain ranges, the Jiulongshuiku site associated with lithic debris and hammerstones would be best interpreted as a chipping station where the first stage of reduction process was carried out.

No site has been found between the four spade-making settlements mentioned above and the oolitic dolomite outcrops, indicating direct access to the raw material by the production communities. The distance of 3-5km between the settlements and outcrops appears to have been the optimal distance for craftsmen to complete several tasks for the raw material procurement within a day. These tasks would have included: walking to the lithic outcrops, quarrying the stone, roughly flaking the stone into preforms, and carrying the preforms home. The fact that our excavations at Huizui have yielded few large flakes supports the scenario that the first stage of manufacture, flaking preforms, may have been completed near the quarries before the raw material reached the tool production sites (Liu et al. 2007). However, due to the lack of dateable material the Jiulongshuiku site has not been dated; therefore, this hypothesis needs to be tested in future excavations.

2.3 Spade production at Huizui

The function of Huizui as a stone tool production locale was first recognised in the 1950s by Chinese archaeologists, based on the discovery of tool blanks at the site (Henan Cultural Bureau 1961). This understanding was then confirmed by our intensive survey of the site in 2000 and subsequent excavations from 2002 to 2006 (Chen 2007; Liu et al. 2002-2004; Ford 2007).

The Huizui site today is separated by a wide gully whose time of formation is unclear, and we refer to the two areas of the site as Huizui East and Huizui West. The Neolithic Yangshao (c. 5000-3000 BC) and late Longshan (c. 2500-2000 BC) people first occupied the site at Huizui East, then the Erlitou (c. 1900-1500 BC) people of the Bronze Age extended their residence to both Huizui East and West, and finally the people of the Eastern Zhou Dynasty (770-221 BC) resided again at Huizui East. The entire site measures about 25 ha, but the occupation area for each cultural phase varied. Intensive surveys and excavations at the site have yielded large quantities of stone blanks and manufacturing debris, and the blanks show different stages of the reduction sequence from raw material to finished products (Figs 4, 6.1-2). These lines of evidence suggest that the site was a stone tool production centre during the Longshan and Erlitou periods (Ford 2004; 2007; Liu et al. 2002-2004).

Our excavations at Huizui have revealed numerous domestic features (such as houses, water wells, burials, and ash pits) filled with animal bones, pottery sherds, charred plant remains, as well as blanks and flakes. This pattern is consistent during the Longshan and Erlitou periods. Our intensive survey, shovel testing and excavation show that stone tool production appears to have been carried out throughout the settlement, according to the sub-surface distribution of lithic debris at the site. We have identified three locations with a high concentration of dolomite and sandstone blanks and numerous flakes from the Erlitou occupation at Huizui West. Based on their depositional forms, the first locale at TP1 was the workplace of a stonemason (Fig. 6.3), the second one at TP2 was a debris dump, perhaps next to a house wall, and the third one at T1H1 was also a debris dump, probably in a courtyard (Ford 2007). TP1 and TP2 appear to have been contemporary (Erlitou phase II), and were spaced about 25m apart, a length perhaps indicating the distance between two households/workshops.

Ninety-four blanks and semi-finished tools collected on the surface at Huizui West during the 2000 field season show a pattern in which oolitic dolomite is the dominant material (78% by number), followed by much smaller proportions of other materials. Oolitic dolomite flakes account for an even higher proportion (92% by weight) of lithic materials from heavy fractions of 111 flotation samples obtained from a pit (T1H1 spits 2-4) in the 2004 excavation. Such high ratios of oolitic dolomite in blanks and flakes are in sharp contrast to the corresponding ratios seen in the finished tool assemblage from the 1959 excavation, in which spades (mostly made of oolitic dolomite) only account for 22% of the major tool types dating to the Erlitou period (Fig. 7) (Liu et al. 2007).

The ratios of tool types and spatial patterns of tool production debris suggest that lithic manufacture occurred within household settings during the Longshan and Erlitou periods; craft specialisation was practiced by a significant number of households; and the major products were oolitic dolomite spades, not only for local use but also for trade and exchange (Chen 2007; Liu et al. 2007).

It is also interesting to note that among the 19 spades excavated in 1959, eight lithic types have been identified, including oolitic dolomite, fine dolomite, siltstone, micaceous fine sandstone, fine sandstone, marble, and lime (burnt). Oolitic dolomite accounts for 47% of the total spades (Table 1). This inventory suggests that various lithic materials were used to make spades which were used at Huizui. This profile is in sharp contrast to the spade blanks and flakes uncovered from the three manufacturing locales at Huizui West, which are almost all identifiable as oolitic dolomite. These phenomena indicate that diverse raw lithic materials available locally were used to make spades for use on site, while mainly one type of raw material was chosen for the intensive production of this type of tool, probably for the purpose of export. This pattern may also suggest that access to oolitic dolomite quarries was controlled by a limited number of families rather than shared by the entire community. This proposition, however, needs to be tested as more data become available.

Table 1: Lithic types of stone spades from Huizui, excavated in 1959

Lithic type Longshan no. (%) Erlitou no. (%) Total no. Total %
oolitic dolomite 2 (25%) 7 (64%) 9 47%
fine dolomite 3 (38%) 1 (9%) 4 21%
siltstone 2 (25%) 0 2 11%
micaceous fine sandstone 0 1 (9%) 1 5%
fine sandstone 0 1 (9%) 1 5%
marble 0 1 (9%) 1 5%
burnt lime 1 (13%) 0 1 5%
total 8 11 19 100%

2.4 Experimental study of spade manufacture

In order to understand the manufacturing processes of stone tools, an experimental study using only raw materials local to the Huizui site was conducted in an attempt to re-create dolomite spades. Five stages of manufacture were identified, including, (1) initial quarrying of stone raw material; (2) crudely shaping/knapping of quarried raw material so as to achieve a size and shape suitable for the next stage of processing (this stage creates the initial preform of the tool); (3) further shaping via finer knapping of edges and ridges, together with hammerdressing to achieve a form that is ready for the next stage; (4) grinding the surface of the tools to further refine the desired shape and size of the tool; and (5) final polishing of the tool (Owen 2007). These results are largely consistent with the previous reconstruction of reduction sequences based on archaeological remains (Fig. 6.1) (Ford 2004).

By mimicking the flake scarring and manufacture/wear patterns identified on the archaeological specimens through experimental knapping, grinding and polishing, a determination of how the originals were produced has been made. However, the polishing stage was not achieved utilising the available Chinese materials. Therefore, further research is needed to determine how such finishes were accomplished in antiquity (Owen 2007).

2.5 Distribution of spades

Oolitic dolomite spades are identified from nearly all the excavated Erlitou culture sites around the Yiluo region, and some of these tools would have been made at Huizui. Among these sites, Zhengyao is located about 100km west of Huizui, and Nanwa is in a different river valley separated from Huizui by the Songshan Mountains (Fig. 8). It is notable that the proportions of oolitic dolomite in spade assemblages from different sites vary significantly. At the Erlitou urban centre, only 24% of spades were made of oolitic dolomite, whereas at Zaojiaoshu, a small site near Erlitou, all the spades in the collection we examined are identifiable as oolitic dolomite (Liu et al. 2007). These data suggest that, as an urban population with rather heterogeneous origins, the Erlitou people obtained tools from various sources. It is also possible that Erlitou, as a political and economic centre, received material and labour tribute from the surrounding regions where various raw materials were used for making spades. In contrast, those small villages far from lithic resources are more likely to have depended on limited lithic production centres in close proximity, such as Huizui, to provide stone products.

The distribution of oolitic dolomite spades and other types of goods in the regional exchange networks is a complex issue, and more trace-element analyses of raw materials and artefacts are needed in order to establish detailed trade patterns.


© Internet Archaeology/Author(s) URL:
Last updated: Wed Jul 29 2009