Introduction

Cultural exchange and interaction have attracted increasing attention1,2,3. During the middle of the Neolithic period, there were frequent social and regional cultural interactions in the middle reaches of the Yangtze River4,5, and the spatial distribution of pottery indicated exchange patterns and interaction processes6,7,8. Pottery embodies the interpersonal connections that continue to be an important element of prehistoric social networks, and the prerequisite for investigating these interpersonal ties is to restore production, distribution and consumption9,10,11,12. The Qujialing culture is one of the most representative archaeological remains in the middle reaches of the Yangtze River. Approximately 3000 BC, the Qujialing Culture expanded to large areas. The Shangang and Gouwan sites had rich cultural remains of the Qujialing culture, but the question of how pottery was disseminated remains unresolved.

Technological analysis is a commonly used method for investigating cross-regional exchanges of pottery13,14,15,16,17,18,19. In this study, the mineral structure, major elements, trace elements and 87Sr/86Sr ratios of samples from the Shangang, Gouwan and Qujialing sites were comprehensively analysed by X-ray diffraction (XRD), inductively coupled plasma optical emission spectrometry (ICP-OES), inductively coupled plasma mass spectrometry (ICP-MS) and Multi-Collector Inductively Coupled Plasma Mass Spectrometer (MC-ICP-MS). These methods can help us understand issues such as technological choices, potter identities, the nature of exchange networks, and pottery provenance20,21.

The paper focuses on 2 questions: were the Qujialing-style pottery from the Gouwan and Shangang sites imported from the Qujialing site, or produced locally? What differences exist between the pottery from the 3 sites?

Methods

Materials

The Qujialing site (30° 50’ 13” N, 112° 54’ 14” E) is located in Qujialing, Hubei Province (Fig. 1). It was the earliest large-scale Neolithic settlement to be discovered and excavated in the middle reaches of the Yangtze River. The site has undergone 4 archaeological excavations, with a total excavated area of 8400 m2. It is a key site for studying the development of civilizations in this region. During the Qujialing culture period, the site became the second largest site in the middle of the Yangtze River22.

The Shangang site (32°57′50″ N, 110°22′30″ E) is located north of Shangang, Xichuan County, Henan (Fig. 1). The site covers 20,000 m2. During the excavation, remains from the Qujialing period, including houses and pits were discovered. The pottery primarily consists of brown sandy ware. Most of the pottery is plain-surfaced, with shapes including Ding-Tripod, Guan-Jar, Bo-Bowl23.

The Gouwan site (33° 05’ 28” N, 111° 28’ 25” E) is located east of Gouwan of Zhangying village, Xichuan County, Henan (Fig. 1). The remains from the Qujialing cultural period at the Gouwan site can be divided into 3 phases: early, middle and late periods24.

Fig. 1
figure 1

Geographical locations of the Gouwan, Shangang and Qujialng sites.

Full size image

Sampling strategy

To determine the composition differences in the Qujialing culture pottery among the Shangang, Gouwan and Qujialing sites as fully as possible, 38 samples were selected from representative excavation units (Fig. 2). On the basis of stratigraphy and typology, these units all belong to the late phase of the Qujialing culture.

Fig. 2: Pottery samples from Shangang, Gouwan and Qujialing sites.
figure 2

a Dou-stemmed bowl (H240: 14) d Bei-cup (H194: 2) Qujialing site: b Dou- stemmed bowl (H5④) Gouwan site: c Ring foot vessel (H87: 5) e Ding-tripod foot (H119: 2) f Ding-tripod (H160 ②: 5).

Full size image

These varieties represent different shapes, colours and functions. To ensure consistency in the experimental conditions, we picked out the admixtures from the coarse-paste pottery under an Olympus SZX12 microscope with 10x.

Mineralogical composition analysis

The experimental setup consists of a D8 Venture X-ray single-crystal diffractometer manufactured by BRUKER, which is equipped with a Cu target and a detector area of 100 × 100 mm. Each sample underwent in situ surface scanning to ensure the accuracy and reproducibility of the experiments while maintaining the sample’s original state as much as possible. After the in-situ surface scanning experiments were completed, the diffraction data were analysed via X-Powder software, and the mineral composition of the samples was determined on the basis of the diffraction patterns.

Chemical composition analysis

The main elements were determined using a HORIBA ULTIMA 2 C ICP-OES, produced by the JOBIN YVON Company. To ensure the uniformity and representativeness of the solid samples, they were ground to less than 200 mesh, digested and tested on the instrument. At least three data points were collected at each measurement location and the average value was taken. The measured elements included 10 major elements: SiO2, Al2O3, CaO, Fe2O3, K2O, MgO, MnO, P2O5, Na2O, and TiO2.

Trace and rare elements were determined by Agilent 7900 ICP-MS by Agilent. It is also required that the sample is ground to less than 200 mesh, and then tested on the machine after digestion. The Agilent 7900 ICP-MS with wet injection was used for full quantitative analysis of samples, applying an external standard with an internal standard (Rh) for accuracy. The accuracy of all elements repeated scanning 5 times reached 1RSD ≤ 5%. There were 41 trace elements, we selected 26 trace elements for in-depth analysis.

Sr isotope analysis

MC-ICP-MS is Thermo Fisher Neptune XT. Firstly, select a fresh surface of the pottery body, grind it to less than 200 mesh, and then use nitric, hydrochloric and hydrofluoric acid for digestion. The digestion solution is passed through Bio-Rad AG50W-X8 cation exchange column to obtain the Sr fraction. A 50 ppb Sr solution is introduced into the Neptune XT MC-ICP-MS for testing the ⁸⁷Sr/⁸⁶Sr isotope ratio using the Cetac Aridus III membrane desorption system. NIST SRM 987 was used as an external standard to correct instrument drift, and 86Sr/88Sr = 0.1194 was used for internal correction of instrument mass fractionation. Standard materials BCR-2 and BHVO-2 were used as quality control blind samples. After chemical digestion, purification, separation, and mass spectrometry testing, their Sr isotope compositions were consistent with the recommended values within the error range.

Results

Mineral compositions

Most pottery is composed of quartz and albite from the Gouwan and Shangang sites, and some samples contain illite, such as SGH138:2, SGH286:11 and GWH371:1. Muscovite appeared in GWH157:7, and dolomite and calcite were found in SGH286:11 (Fig. 3a). As shown in Fig. 3b, the Qujialing samples are composed of quartz, albite and illite. The mineral composition of the pottery varied both between different sites and within the same site.

Fig. 3: X-ray diffraction pattern.
figure 3

a Mineralogical composition analysis diagram of Gouwan and Shangang sites; b Mineralogical composition analysis diagram of Qujialing site.

Full size image

Major elements

Ancient clay is generally divided into 4 categories according to the contents of SiO2, and Al2O3 and the flux. These categories include: ordinary clay, Mg-rich, Si-rich and Al-rich clay25. The SiO2 content at the Gouwan site is mostly 56.32–71.09%, the Al2O3 content is 13.29–17.32%, and the flux content (∑flux = CaO + Fe2O3 + K2O + MgO + MnO + Na2O + P2O5 + TiO2) is mostly 12.55–19.06%; the SiO2, Al2O3 and flux contents range from 58.11% to 69.21%, 14.17% to 18.72% and 12.68% to 17.53% respectively at the Shangang site; the SiO2 content in the Qujialing site is 61.23–71.09%, the Al2O3 content is 12.91–17.71%, and the flux is 12.59–15.74%.

Selecting the raw materials is critical to ensuring the quality and performance of pottery products. The choice of raw materials results from complex interaction among its available within the local landscape, the optimization of technological processes, and the need for specific materials qualities26,27,28. To more clearly show the differences in the major elements, we conducted a statistical analysis, and the SiO2, Al2O3, Fe2O3 and MnO contents did not differ. Differences in other major elements are shown in Fig. 4; significant differences in terms of CaO, K2O, Na2O, P2O5, TiO2, and MgO exist between the Gouwan and Qujialing sites, as well as between the Shangang and Qujialing sites in CaO, Na2O, TiO2, and MgO. Furthermore, there are significant differences in K2O, P2O5, and MgO at the Gouwan and Shangang sites (Fig 5).

Fig. 4
figure 4

Plot of the principal components.

Full size image
Fig. 5: Major elements difference plot.
figure 5

a–f Box and whisker plot variations in CaO, K2O, MgO, Na2O, P2O and TiO2 content; * significant difference; ** highly significant difference.

Full size image

The results of the KMO and Bartlett analyses revealed that KMO = 0.643, P <0.01, indicating that the principal and component analyses were effective. Factor analysis revealed 3 major components, accounting for 80.56% of the total variance. Of these, F1 explained 38.2% of the variance, which was the highest contribution rate (Fig. 5). It was dominated by MgO (load = 0.95) and Al2O3 (load = 0.82).

A principal component analysis revealed that the Shangang site exhibited clear clustering, whereas the Gouwan and Qujialing sites presented more scattered data distributions. The differences in major elements and PCA results inferred that there were different pottery recipes at the 3 sites.

Cluster analysis was conducted on the major elements present in 38 samples, and the results are shown in Fig. 6. Initially, the samples were divided into 2 categories. H437:2 from the Gouwan site is an Al-rich pottery that does not cluster with other ordinary pottery samples. This finding is consistent with the chemical composition analysis. The secondary cluster indicates that the pottery from the Qujialing site constitutes an independent group. The remaining pottery samples were categorized into Groups 3, 4, and 5. Compared with Groups 4 and 5, Group 3 contains a greater proportion of Al2O3. Group 4 can be designated the Shangang group, distinguished by its lower silicon content. Group 5 is the Gouwan group, which has a relatively high silicon content. Although there are different load elements in PCA and CA, the data are distinct.

Fig. 6
figure 6

Hierarchical cluster analysis diagram of major elements.

Full size image

Trace (rare) elements

In the analysis of pottery production, the use of trace element analysis is more effective29,30,31,32,33. Trace elements preserve the natural regional characteristics of clay and can be used to assist in identifying the soil resources used in pottery production and potential production sites34,35,36. After standardizing the trace elements and plotting a point-line diagram37,38, obvious differences in the Ba and P contents of the pottery samples from the 3 sites were observed (Fig. 7a). Figure 7b shows the elemental distribution patterns of the samples. The distribution curves of the elemental distribution patterns in the pottery are similar, all exhibiting a pattern of enrichment in light rare earth elements and depletion in heavy rare earth elements, with negative anomalies in the Eu values.

Fig. 7: Trace elements line chart.
figure 7

Line graphs showing trace elements (a) and rare earth elements (b) (ppm).

Full size image

GWH87:5, and GWH157:2 presented positive Ce values, which distinguishes them from other potteries at the site. This characteristic reveals the diversity of raw materials used in pottery at the Gouwan site. Pottery samples from the Qujialing site showed positive Yb values, whereas those from the Gouwan and Shangang sites exhibited negative Yb. This discrepancy confirms the different sources.

In different types of soil samples, the concentrations of 12 trace elements—Ba, Co, Cr, Cs, Hf, Nb, Ni, Rb, Sr, Th, V, and Zr—remain constant during the pottery firing process. Therefore, these elements can be used as characteristic indicators for provenance analysis39. Owning to this stability, we reduced the dimensions of the 12 trace elements and plotted a scatter diagram. As shown in Fig. 8, although the pottery samples from the 3 sites exhibited consistent cultural styles, they presented distinct distribution patterns in the terms of the compositions of stable trace elements. The stable trace element concentrations from the Shangang and Gouwan sites show similar distribution patterns, possibly due to their proximity in geographical locations. The Qujialing site exhibited the greatest dispersion, suggesting more diverse clay sources.

Fig. 8
figure 8

Plot of the stable trace elements.

Full size image

87Sr/86Sr ratios

Establishing the provenance of exotic materials and identifying the trade networks through which the materials moved remain central to larger interpretations and debates in archaeology16. In recent years, many researchers have used strontium isotopes, which can be used to determine the provenance of geomaterials40. Sr isotopes have become key elements in tracing ancient humans41,42, animal migration43 and sources of pottery44,45,46,47.

A total of 17 samples from the Shangang site had 87Sr/86Sr ratios ranging from 0.713653 to 0.717139, with an average of 0.714760, and a median of 0.714514, indicating a concentrated distribution. The distribution range of 87Sr/86Sr for the pottery samples from the Gouwan site was 0.713823–0.721358, with an average value of 0.717056 and a median value of 0.716807, indicating a dispersed distribution. Seven pottery samples from the Qujialing site yield 87Sr/86Sr values of 0.714678–0.722869, with an average of 0.717961 and a median of 0.717577, demonstrating the greatest dispersion.

The 3 sites are situated on the Han River, which is the largest tributary of the Yangtze River. The 87Sr/86Sr ratio of groundwater can serve as an effective indicator of interactions between various rock minerals40,48. According to Yang’s study of Sr isotopes in the Yangtze River, the 87Sr/86Sr ratio of suspended solids in the Han River- Xiantao section is 0.721143 ± 3, and that of the sediment is 0.713117 ± 13. The 87Sr/86Sr of suspended solids in the Han River- Wuhan section is 0.723963 ± 349. We use these 3 sets of data as the geochemical Sr endmembers for the Han River. The 87Sr/86Sr ratios of the pottery samples fall within this range (Fig. 9).

Verifying the assumption that the samples are of the same provenance can be achieved by confirming that all the data align with the same isochron line50. As Fig. 10 illustrates, independent fitting lines were formed at the Shangang, Gouwan and Qujialing sites, with respective goodness-of-fit values of 0.70, 0.87 and 0.91. Strong correlations were observed among the sites. ANOVA is a helpful statistical method for determining whether there are significant differences between groups51. The results indicate that there are significant differences in the 87Sr/86Sr ratios among the 3 sites (P < 0.001).

Fig. 9
figure 9

A comparison of Sr isotope data among 3 sites from Qujialing culture.

Full size image
Fig. 10
figure 10

Box and whisker plot of the 87Sr/86Sr.

Full size image

The Sr isotope analysis suggested that the pottery could be produced and consumed within the sites.

Discussion

To investigate the production and usage patterns of pottery of the same cultural style within an independent space, chemical composition, mineral structure, and Sr isotope analyses were conducted on 38 pieces of pottery from the Shangang, Gouwan and Qujialing sites.

In terms of mineral phases, the samples from the Qujialing site are primarily composed of quartz, albite, and illite; those from the Shangang site include quartz, albite, illite, calcite and dolomite; and those from the Gouwan site are predominantly composed of quartz, albite, illite, and muscovite.

The Shangang site yielded 2 types of clay: ordinary and high-silica clay. The Gouwan site included 3 types of clay: ordinary, high-silica and Al-rich clay. All the pottery from the Qujialing site was made from ordinary clay. PCA and CA indicated diverse selections of clay, representing different pottery recipes, among the 3 sites.

The consistency of major elements is an important factor in measuring the provenance of pottery52,53,54. Independent sample t-tests, PCA and CA revealed that the major elements in the pottery samples from the 3 sites were significantly different and were relatively independently clustered. These discrepancies were the result of conscious technical choices made by the potters during the pottery-making process.

According to previous research, the pottery from the Zoumaling site of the Qujialing culture consists of common, high-silica and Al-rich clay, and that from the Fenghuangzui site has 2 types of clay: high-silica, and high-aluminium. Moreover, the pottery found at these 2 sites should be produced locally55,56. The eggshell-black pottery from the Qujialing site is made from Al-rich clay, and the mineral phases include orthoclase, hercynite, etc.57. These differences indicate that within the distribution range of the Qujialing culture, the pottery from different sites exhibited distinct formulations.

The 87Sr/86Sr analysis revealed significant differences, and the trace elements Ba and P, and the rare element Yb varied significantly among the Shangang, Gouwan and Qujialing sites. We argued that each site produced, and used pottery independently, and that communication within the Qujialing culture relied mainly on people’s movements.

The analysis facilitates a more comprehensive understanding of the production, and use of pottery, and provides scientific data support for restoring a model of prehistoric cultural exchange and interaction in the Yangtze River Basin.

Owing to the limited number of samples, we were unable to test and analyse the pottery of different textures and shapes found at each site. Further archaeological materials and more accurate, diverse strontium isotope data are needed.