Sourcing celadons with EDXRF and LA-ICP-MS from the Xunyang city burial complex, 202 B.C–907 A.D.

Abstract

This paper uses a combination of EDXRF and LA-ICP-MS to analyse the chemical components including major elements and REE composition of the accompanying celadon from four different periods of tomb at the Yutu Hill burial complex on the Xunyang city site for sourcing their provenance. The combination of the two analytical methods not only provides additional data and information to explore the provenance of the accompanying celadon in the burials, but also validates the data’s accuracy. A comparison with the chemical composition of the products from several kiln sites producing celadon revealed that burials from the Eastern Han period (25–220 A.D) to the Tang dynasty (618–907 A.D) included celadon from the Hongzhou kiln, indicating the connection between the Xunyang city site and the Hongzhou kiln during these two periods. In addition, the REE composition of the accompanying celadon in tombs from the Six Dynasties (222–589 A.D) and the Sui dynasty (581–618 A.D) is very similar to that of Yue kiln celadon, despite their strikingly different appearance and shape. This paper deduces, based on the previous distribution of Yue kiln products in the Jiangxi region, that only the upper classes collected Yue kiln products during the initial period of their entry into Jiangxi during the Jin dynasty (266–420 A.D), but that during the later period of the Six Dynasties, the commoner classes also began to collect Yue kiln products.

Introduction

Sourcing burial objects from the tombs of large sites can aid in understanding the site’s past trading relationship with the surrounding kiln sites and the site's history as a whole. In the past, the provenance and date of burial ceramic objects were determined using typological methods, but for objects with multiple potential provenances, such as Qingbai ware (a typical porcelain products which were produced by more than fifteen kiln sites in different regions in China), it is difficult to determine the provenance of the objects in terms of a typological approach [1–3]. With the introduction of more elemental techniques into archaeological research, the tracing of burial ceramics and export ceramics can be based on the unique geochemical composition of pottery artefacts combined with factor analysis of various chemical composition to determine their possible provenance [4–7]. Many studies [3, 8–11] in the past have successfully aided in the tracing of ceramic products from burials or exports by employing both main elements and REE components. Xu et al. [9] used REE components to trace the geochemical characteristics of Qingbai ware from the shipwreck in the Java Sea, revealing that the majority of the Qingbai ware originated from various kilns in the Fujian region. Zhu et al. [3] used major chemical and REE components to trace the ceramics excavated from the Xicun region and found that Jingdezhen products unearthed from the Xicun area in Guangdong region due to the comparison of peculiar geological characteristics. Li et al. [21] used REE components to successfully distinguish the Qingbai ware excavated from the Husi kiln from other kilns. These findings support the use of primary elemental and geochemical composition comparisons in tracing ceramic artefacts excavated from burials and shipwrecks.

The subject of this paper is the burial celadon excavated from the Yutu Hill burial complex at the Xunyang city site which is one of the most important sites in the Jiangxi region. From the Han dynasty (202 B.C–220 A.D) to the Tang dynasty (618–907 A.D), celadon was a dominant component in the production of Chinese ceramics [12–14]. During this time, kiln sites in the majority of China's regions, including those along the Hong River, were influenced by the Yue kilns to produce celadon products similar to those of the Yue kilns (The territory that is now Vietnam was once part of China and produced celadons during the relevant time periods) [15–17]. With the exception of the high-quality Yue kiln celadon, many kiln sites produced ordinary civilian celadon that is difficult to distinguish based on appearance alone [18]. In China, celadon was produced at over twenty large kiln sites, including the Dingzhou kiln in Shaanxi, the Yue and Wuzhou kilns in Zhejiang, the Hongzhou and Yuezhou kilns in Jiangxi, the Shouzhou kiln in Anhui, and the Gongxian kiln in Henan before the Song dynasty (960–1279 A.D).

Previous comparative studies [13, 16, 19–21] on identical-appearing wares from these kiln sites producing celadon have been conducted in the past, and the celadon products from these various kiln sites can be distinguished based on their REE composition and the comparison of the principal chemical elements by factor analysis of their chemical composition. Some researchers [19, 20] even have conducted factor analyses of the chemical composition of celadon products of varying ages from the same kiln site to identify differences between products of varying ages from the same kiln site. By analysing the REE composition of celadon excavated from Yingou kiln in Fuping area (Dingzhou kiln), Liu et al. [21] were able to distinguish between celadon from Hongzhou kiln from the Eastern Han, Sixth Dynasty, Sui, and Tang dynasties and celadon excavated from Dingzhou kiln in Fuping area. Through factor analysis of REE components, Feng et al. [20] revealed the distinction between Yue kiln and Hongzhou kiln-produced celadon, and they also compared celadon products from different eras of the Hongzhou kiln and discovered that they possessed distinct geochemical characteristics. These previous chemical analyses of Chinese celadon have provided a solid foundation for tracing the burial celadon unearthed from the Yutu Hill burial complex at the Xunyang city site. This paper will combine EDXRF and LA-ICP-MS techniques to analyse both the major and trace elements of the burial celadons at the Xunyang city site. After obtaining the chemical composition of these burial celadons, the paper will be compared to the celadon data excavated from these kiln sites in order to determine the origin of these burial celadons.

Archaeological background

Located on the western shore of Baili Lake in present-day Jiujiang city, the Xunyang city site is one of the most crucial culture remains for researching the past developments of the middle of Yangtze River region. The ruins of Xunyang city were discovered in the 1980s, and on two separate occasions in 1981 and 1985, the cultural relics management department of Jiangxi Province and provincial and municipal cultural relics experts focused on fieldwork and archaeological identification of the site [22]. The inspection revealed that there were relics and remnants of the ancient city in the vicinity of Maanzhou, Weizui, Yubu Mountain, Baitai Mountain, Qili Lake, and Hetianxai in the village of Saihu, spanning an area of approximately 3 square kilometres. Archaeological evidence identifies the site as Xunyang City crossing six dynasties. Xunyang city site spans a very long period of time, being one of the most prosperous towns in Jiangxi region from the Han Dynasty through to the Tang Dynasty for over 1000 years [23].

In 2019, a burial complex was discovered in the Yutu Hill region of the Xunyang City site, with an excavated area of 2615 square metres (Fig. 1). Based on the area of existing exposures, a total of 23 burials were discovered, including 19 burials with early celadon, three with late celadon, and one with Jizhou ware of the Yuan dynasty. The burials contained an abundance of cultural artefacts, such as pottery, silver, bronze, and gold ornaments.

Fig. 1

Detailed map of the excavation of the burial complex on Yutu Hill at the Xunyang city site. (Map made by the authors using ArcGIS Pro)

An examination of the coins recovered from the graves revealed that the earliest grave, M1, dates to the Eastern Han period in terms of the discovery of ‘Wuzhu’ and ‘Daquan’ coins. The burial M16 yielded Northern Song dynasty ‘Zhihe’ currency and jizhou ware with a Yuan dynasty commemorative mark. Thus, the burial group spans from the Eastern Han to the Yuan dynasties. The latest results of excavations in this area indicate that Yutu Hill, as a burial area, spans a long period of time from the Western Han period to the Ming and Qing dynasties, and is an important site for an ancient tomb group.

As Table 1 shown, the scale of the burials and the varying number of burial objects suggest that the identity of the deceased should have been differentiated between the tombs. M8 is the highest ranking of the early tombs in which celadon was excavated, with a sloping pathway and a white plaster mud base, a form of burial found in the Western Han tombs of Changsha, a relic of the Chu culture. A personal seal and a bronze tripod were unearthed from M14, which suggests that the identity of the tomb owner was a general. This portends a complex admixture of burial groups from the lower and upper classes that may reflect the evolution of the Xunyang city site over time. The variations in burial objects reflect the differences in class as well. Understanding the trade relations between the Xunyang city site and the nearby kiln sites at that time can be done by studying the various burial celadon periods.

Table 1 Detailed description of the Yutu Hill burial complex at the Xunyang city site

The burial objects from four well-preserved tombs of different periods are selected for analysis: M12 from the Tang dynasty (618–907 A.D), M15 from the Sui dynasty (581–618 A.D), M19 from the Six Dynasties (222–589 A.D) and M21 from the Eastern Han dynasty (25–220 A.D). The burial celadon excavated from these four tombs is relatively complete, and the burial forms vary, with the characteristics of their corresponding periods. In terms of stratigraphic relationships, the four burials were not affected or overlaid by the stratigraphy of other burials, and there is no possibility of the upper layer of funerary objects falling to the lower layer.

Material and methodology

Samples

This study collected seven burial celadon samples from four tombs of Yutu Hill burial complex. One each from the M12, three from the M15, one each from the M19 and M21. The accompanying celadon from M12 is an ewer (Fig. 2a), labelled M12, of which the spout is broken (Fig. 2b), and a sample was taken from the broken spout for analysis in this study. Three bowls with the designations M15-1, M15-2, and M15-3 are the three-accompanying celadons from M15. They each have the designations M15-1, M15-2, and M15-3. M15-1 is a footless bowl with a heavily flaked surface glaze that was discovered in fragments (Fig. 2c). M15-2 is a footless bowl with a brown glaze and an interior floral carving that is in good condition (Fig. 2d). M15-3 is a small bowl with a foot and a light brown glaze that has been preserved well (Fig. 2e). A jar with a lid marked M19 that contains the accompanying celadon has very heavy glaze flaking and a broken lid; samples were taken from the broken lid for this study's analysis (Fig. 2g, j). A small jar labelled M21 that contained the accompanying celadon was discovered broken into two pieces, and samples were taken from one of these parts for this study's analysis (Fig. 2f). It needs to be specified that the ceramic samples chosen for this study are of the broad celadon type. They differ from the conventional definition of celadon in that they are yellow rather than green and of an earlier celadon variety.

Fig. 2

Photos and dimensions of the samples analysed in this study

Analysis approaches

EDXRF

The major and minor chemical composition analysis for the burial celadons were analysed by an energy-dispersive X-ray fluorescence technique and an EDAX Eagle III XXL spectrometer, the quantitative chemical compositions of glazes on these specimens and raw materials were analysed (produced by the EDAX unit of AMETEK, Inc., USA). The utilised EDAX EAGLE III features a 40 W (40 kV, 1000 A) X-ray tube and a Si (Li) detector. The testing mode was surface scan, and the diameter of the X-ray focus was 300 m. The counting duration was 600 s. High voltage and low current (50 kV, 200 A) was used to analyse the content of MnO and Fe2O3, whereas low voltage and high current (10 kV, 700 A) was used to analyse the content of Na2O, MgO, Al2O3, SiO2, P2O5, K2O, CaO, and TiO2. For the analysis of samples of botanic ash and glaze ash, pellets formed by pressing 5 g of ash with a force of 150 kN were used.

For quantitatively calculating the results of an analysis, standard curves were utilised. The calibration curves were established using the set of ten ancient ceramic reference materials developed by the Shanghai Ceramic Institute of the Chinese Academy of Sciences. The analysis was monitored for precision and accuracy using GBW07402 (clay), a national standard material. GBW07402 was analysed nine times throughout the course of the analysis of the 54 samples. The relative standard deviation (RSD) of the ten reported oxide concentrations was 5%. The average of the nine analytical results was in good agreement with the recommended chemical composition of GBW07402: the relative error of the nine reported oxides was 5% and the relative error of P2O5 was 10%.

Additional file 1: Table S1 in displays the average quantitative chemical composition of 10 major and minor elemental oxides for each sample group, as well as the standard deviation (SD) of individual results within each group.

LA-ICP-MS

In this study, the REE composition of the burial celadons from tombs was analysed using LA-ICP-MS. The REE composition was selected as the markers for geochemical analysis of burial ceramics collected for this study because the main elements of the products produced in the areas where early celadon was produced, such as Yue, Yuezhou, and Hongzhou kilns, are very similar. All the samples was conducted by Laser Ablation Inductively Coupled Plasma Mass Spectrometry (LA-ICP-MS). The pre-treatment and LA-ICP-MS analysis of the Husi wares were performed at the CAS Key laboratory of Crust-Mantle Materials and Environments, University of Science and Technology of China. The analysis was undertaken using a Coherent ArF excimer UV 193 nm wavelength Laser ablation system (GeoLas pro) coupled with Agilent 7700e quadrupole ICP-MS.

The samples were cut using a 0.8 mm-thick diamond wire cutter and polishing the cross-section of the specimens to preserve both the clear glaze layer and the clay body layer. The 38 μm spot size and a 10 Hz ablation frequency were used for single spot ablation. The chemical compositional data were calculated via “multi-external references without internal standards calculating methods of Silicate” provided by ICP-MSDataCal [24, 25]. The reference materials consist of NIST SRM 610 and NIST SRM 612 from the National Institute of Standards and Technology, as well as the USGS (United States Geological Survey) reference glasses BCR‐2G, BHVO‐2G and BIR‐1G. The NIST SRM 610 was used as a monitoring standard. The BCR‐2G, BHVO‐2G and BIR‐1G were used as external calibration standards, and the NIST SRM 612 was used as a secondary standard [26]. The accuracy and precision, including absolute error, relative error, standard deviation and relative standard deviation achieved by the analyses of the secondary standard reference NIST SRM 612, for this study are provided in Additional file 1: Table S4. All quantitative methods utilised in this paper are Exploratory Factor Analysis, in which multiple variables are downscaled in order to calculate their factor scores, and then the differences between samples are examined [27].

Statistic methods

Statistical Product and Service Solutions (SPSS) software was utilised for all analyses in this report. We utilise SPSS Statistics 29.0.0 as our software version. This paper utilises factor analysis to reduce the dimensionality of the data. The extraction method for factor analysis is Maximum likelihood, while the rotation method is Varimax. This paper utilised these extraction and rotation method because the vast majority of prior research [18, 19, 32] utilised these extraction and rotation techniques for factor analysis. In order to facilitate a more accurate comparison with their data, the same methods of analysis were used in this study.

Meanwhile, KMO and Bartlett's test for sphericity was used to evaluate the suitability of the data for statistical analysis in this paper. The Kaiser–Meyer–Olkin Measure of Sampling Adequacy is a method for determining whether or not data is suitable for Factor Analysis, an indicator used to compare simple and biassed correlation coefficients between variables [36]. When the sum of squares of simple correlation coefficients between all variables is significantly greater than that of partial correlation coefficients. The closer the KMO value is to 1, the greater the correlation between variables and the greater the suitability of the original variables for factor analysis. The closer the KMO value is to zero, the lower the correlation between variables and the less suitable the original variables are for factor analysis [36]. As a result, KMO data are provided in all analysis plots throughout this paper.

Results and discussion

To facilitate the presentation of statistical results, each sample is labelled with the number of the grave in which it was discovered. The results of the EDXRF and LA-ICP-MS analyses are largely comparable (Additional file 1: Tables S1, S2 and S3), confirming the accuracy of the chemical composition results for the accompanying celadon. The primary components of the M12 and M19 clay bodies are significantly different from those of the M15 and M21 bodies. As Fig. 3a shown, the aluminium oxide and potassium oxide contents of the M12 and M19 bodies are significantly lower than those of the M15 and M21 bodies (Al₂O₃: average 16.5% < 22.0%; K₂O: average 1.9% < 3.9%), whereas the silicon oxide content is greater than that of the M15 and M21 bodies (average 76.4% > 67.2%). According to the primary composition of the carcass, it would appear that M15 and M21 are not of the same origin as M15 and M21. In addition, from a geochemical point of view, the Nb/Ta and Zr/Hf ratios of M12 and M19 are very different from those of M15 and M21 (Fig. 3b), which reinforces the fact that the burial celadon of M12 and M19 is of a different provenance from that of M15 and M21. In terms of both factor analysis and comparison of REE composition, M12 and M19 should have the same provenance. M15 and M21 would have belonged to a different provenance.

Fig. 3

a SiO₂/Al₂O₃–K₂O scatter plot of the clay body of the burial celadons; b Nb/Ta–Zr/Hf scatter plot of the clay body of the burial celadons

Having discovered that they are of different origins, tracing them poses an additional formidable challenge. This study gathered information on kiln sites that produced similar celadon, such as the Yuezhou and Hongzhou kilns. Unfortunately, none of the three sites have sufficient data for the direct geochemical analysis of Nb, Ta, and Zr from the clay source. Fortunately, Zhu et al. [16] have been conducted to differentiate between the three kiln sites based on other REE components of clay body, including Rb, Co, Cs, and Ba. The difference between the Rb values of the Yue and Hongzhou kilns is enormous. Nevertheless, there is no link to any particular minerals or rocks for the presence of rubidium because this element is always present as an impurity in sodium and potassium. On the other hand, The potassium oxide content of the clay bodies of the burial wares from M12 and M19 is significantly different from that of the Hongzhou kiln product and more comparable to that of the Yue kiln product.. Moreover, on the basis of Zhu et al. [16]’s the Rb and Hf data of the accompanying celadon from M12, M15, M19, and M21 were substituted into the analysis. The Rb-Hf of M12 and M19 fell within the interval of the Yue kilns, whereas M15 and M21 fell within the interval of the Hongzhou kilns (Fig. 4a, b). Moreover, according to the factor analysis (variables Rb, Hf, Co, Cs and K) of Zhu et al. [16], the accompanying celadons from M15 and M21 fall into the Hongzhou kiln area, while those from M12 and M19 fall into the Yue kiln area (Fig. 4c, KMO = 0.77). In conjunction with the previous analyses of Nb/Ta and Zr/Hf, it is possible to conclude that the accompanying celadon of M12 and M19 originated from Yue kilns, while that of M15 and M21 originated from Hongzhou kilns.

Fig. 4

a and b: Rb-Hf scatter plot of the clay body of analysed samples and the celadon from Hongzhou kiln and Yue kiln; c Exploratory Factor analysis (KMO = 0.77) for clay body of all the samples analysed in this study compared with data from previous studies [16, 19, 20, 31] (the cited data is compiled in Additional file 2: Tables S5–S7)

Chronology of the burial celadons from M15 and M21

Using Ba, Co, Fe, Th, U, Rb, and Na as independent variables, Feng et al. [20] ran an Exploratory Factor Analysis for the chemical components on the various dates at which objects were fired in the Hongzhou kilns. It was discovered that the factor analysis results for these REE component variables could be used as a basis for determining the age of the accompanying celadon from the Hongzhou kiln in the burial, and that the direct factor analysis results for the artefacts excavated from the Hongzhou kiln varied considerably between the different dates.

Based on the study by Feng et al. [20] (the analytical data can be found in Table S5–S7 of the Additional file 2), factor analysis was conducted by substituting Ba, Co, Fe, Th, U, Rb, and Na for M15 and M21 as variables. As Fig. 5 (KMO = 0.81) shown, the celadon from M15 is from the same region as the celadon from the Eastern Han period excavated at the Hongzhou kiln, and the celadon from M21 falls in the area between the Sui and Tang dynasties’ data of the Hongzhou kiln. In conjunction with the results of the previous Rb-Hf analysis, the accompanying celadon excavated from M15 and M21 dates to the Eastern Han period and the Sui or Tang periods, respectively. Chemical composition comparison reveals that the chronology of the burial celadon excavated from M15 and M21 is more consistent with that of their respective burials. Therefore, the accompanying celadon from M15 and M21 can be identified as a product of the Hongzhou kiln.

Fig. 5

Exploratory Factor analysis (KMO = 0.81) for the clay body of samples from M15 and M21 tombs analysed in this study compared with data from previous studies of the Hongzhou kiln [20] (the cited data is compiled in Additional file 2: Table S7)

Chronology of the burial celadons from M12 and M19

A comparison of the main volume and REE composition of the accompanying wares revealed that the celadon from burials M12 and M19 was from an entirely different source than that from burials M15 and M21. Specifically, from a geochemical standpoint, the Nb/Ta and Zr/Hf ratios predict a substantial difference in the origin of the clay used to fire the celadon. In addition, the Rb values for Hongzhou kilns in the Jiangxi region are above 240, whereas the celadon from M12 and M19 has an average Rb value of only 121. These indicate that the accompanying wares from M12 and M19 are not likely to have originated from the Hongzhou kiln in Jiangxi region but Yue kiln in Zhejiang region.

On the other hand, neither M12 nor M19 are consistent with Yue kiln in terms of the appearance and shape of the celadon, particularly the body colour. The Yue kiln were renowned for the quality of their firing, the body's tightness, and the number of washing cycles, which resulted in low iron content and a white body [12]. M12 and M19, on the other hand, are coarser and grayer, indicating less frequent washing and a higher iron concentration in the clay bodies. Comparing the production process and the celadon unearthed from previous Yue kilns, it is evident that M12 and M19 were not produced at Yue kilns. However, the REE composition reveal significant differences between them and the Hongzhou kilns, further confounding the origin of the accompanying celadon excavated from M12 and M19.

The chemical compositions of other well-known kiln sites producing celadon, including Dingzhou kiln, Shouzhou kiln, Wuzhou kiln, Yaozhou kiln, and Yuezhou kiln [16, 19, 21, 28–33], were collected for comparison in this study in order to determine the origin of the accompanying celadon excavated from M12 and M19. Na, Mg, Si, Al, K, Ca, and Ti were used as the seven primary variables for the factor analysis. As shown in Fig. 6 (KMO = 0.73), all of these kiln sites differ significantly from the areas of the accompanying celadon excavated at M12 and M19 in terms of factor scores. Complementary celadon from M12 and M19 does not match the compositional profile of these renowned celadon kiln sites. In relation to the celadon of each kiln site, the major chemical and REE compositions of M12 and M19 are still considerably closer to those of the Yue kilns.

Fig. 6

Exploratory Factor analysis (KMO = 0.73) for the clay body of samples from M12 and M19 tombs analysed in this study compared with data from previous studies [16, 19, 21, 28–33] (the cited data is compiled in Additional file 2: Tables S5–S14)

The accompanying celadon unearthed from M12 and M19 differs significantly in appearance and shape from that of the Yue kilns. However, in terms of REE and major elements composition of the clay bodies, M12 and M19 are more similar to those from Yue kilns. Assuming that the accompanying celadon excavated from M12 and M19 is the products from Yue kiln, a previous study [16, 19] analysed the composition of Yue kiln celadon from various eras. Based on the chemical composition data of the glaze, using the same seven elements as above as variables for factor analysis, the burial celadon from M19 is more similar to the Wapiantan kiln site of Yue kiln in the Five dynasties (Fig. 7: KMO = 0.70). The glaze chemistry of celadon from the Yue kiln site of the Sixth Dynasty is very similar to the burial celadon from the M12 (Fig. 7). On the assumption that M12 and M19 are products of the Yue kiln, a comparison of their glazes with those of other Yue kiln products from different time periods reveals that the dates of the accompanying celadon from M12 and M19 can correspond to the dates of their tombs. This provides additional evidence that the burial celadon of M12 and M19 are Yue kiln products despite their significantly different external appearances. In contrast, the primary component of the glaze of M12 is significantly different from that of a Yue kiln product.

Fig. 7

Exploratory Factor analysis (KMO = 0.70) for the glaze of samples from M12 and M19 tombs analysed in this study compared with data from previous studies of Yue kiln [16, 19, 21, 28–32] (the cited data is compiled in Additional file 2: Tables S10 and S11)

Typology of studied samples of relevant kiln sites

According to the results of the data analysis and comparison, the Hongzhou kiln and the Yue kiln are associated with the burial celadon of the Yutu mountain burial complex at the Xunyang city site. According to Zhang [34], many opinions believe that all celadon produced at Yue kilns was green, but during the early period, Yue kilns also produced yellowish glazes. Wood [12] observed that the yellowish-glazed celadon wares produced by Yue kiln used the same glaze formula as the green-glazed wares, with the difference in colour resulting from the different firing atmosphere. Iron oxide in the glaze is converted to ferrous oxide during reducing firing, which is the primary cause of the glaze's greenish hue. In the absence of ferrous oxide, the glaze will be yellow if it is fired in an oxidising atmosphere. Glaze colour cannot therefore be used to determine the relationship between these burial complexes and the Yue kiln. On the other hand, the early Yue kiln celadon products rarely exhibit significant flaking, unlike the accompanying celadons from the Yutu mountain burial group, which exhibit significant flaking. This indicates that the difference between the coefficient of expansion of the glaze and the body of the funerary celadons is too great, resulting in traces of significant glaze flaking, a phenomenon frequently observed in Hongzhou kiln wares [32].

The greatest difference between the accompanying wares and Yue kiln products is reflected in the clay body. Yue kiln is known for its excellent clay-processing technology, which resulting in its clay body very tight carcass structure, low sand inclusions and a carcass colour that rarely appears greyish-yellow [34]. However, all of the funerary objects excavated from the Yutu mountain burial group have a loose body structure, with a high number of sand inclusions, and most have a greyish-yellow colour, which is very similar to the clay body of the Hongzhou kiln product [20]. In addition, Yue kiln celadons are often accompanied by decoration such as rope motifs and carvings [34], the Yutu mountain burial mound has very little decoration on the accompanying wares, which is similar to that of the Hongzhou kiln.

On the basis of a comparison of the appearance of the celadons, the four burial sites in the Yutu Mountain burial complex, M12, M15, M19, and M21, have similar body and decoration characteristics to those of the Hongzhou kiln and differ significantly from those of the Yue kiln. In contrast, the major chemical and REE compositions of M12 and M19 are significantly different from those of the Hongzhou kilns and are more comparable to those of the Yue kiln. These are the two hypotheses derived from this study:

• 1. The burial celadons from Xunyang city site are produced in the Yue kiln region, but not at the main Yue kiln site. Rather, they are produced at common kiln locations dispersed throughout the region. Therefore, the REE components of the body are comparable to those of the Yue kiln.

• 2. The concept that each elementary signature is uniquely linked to a site type (kiln) requires additional research. In reality, it depends on the availability of raw materials and the geological distribution of rocks. Thus, it has been demonstrated that sites from different locations, but distributed along the same river and utilising the same river sand/mud, had comparable signatures of the major chemical constituents [35].

The accompanying wares from M12 and M19 are not only similar to Yue kiln products in terms of their primary elements, but also in terms of their REE composition. Therefore, it is more likely that the first of these hypotheses was formulated at a kiln site surrounding Yue kiln, as opposed to the main sites of Yue kiln.

Trade ties between the Xunyang city site and neighbouring kiln sites

As mentioned above, the accompanying celadons from M15 and M21 are consistent with Hongzhou kiln celadon products of their respective periods in terms of both form and chemical composition. M15 was more similar to Hongzhou kilns of the Eastern Han period, while M21 was more consistent with the REE composition of Hongzhou kilns from the Sui and Tang periods in terms of a previous factor analysis of the REE composition. This indicates that trade between Xunyang City and the Hongzhou kilns began during the Eastern Han period and continued through the Sui and Tang dynasties. Previous research has shown that the Hongzhou kilns began to decline during the Five Dynasties [20], it is likely that trade between Xunyang City and the Hongzhou kilns continued until their firing ceased. Ancient Xunyang City was 192 km away from the Hongzhou kilns, according to the map (Fig. 8), and there were several waterways around the kilns that led directly to Xunyang City, which may have been the basis of trade between the two cities for hundreds of years.

Fig. 8

Relationship diagram between the Xunyang city site and the locations of the Hongzhou and Yue kilns

Compared to Hongzhou kilns, the geochemical composition of the accompanying celadons unearthed from burials M12 and M19 are much more similar to that of Yue kiln products. In addition, the chemical composition of the glazes of the accompanying celadon excavated from M12 and M19 corresponds to different periods of Yue kiln products, and the corresponding dates correspond to those of the burials, indicating that the accompanying celadon from M12 and M19 is likely a product of the Yue kiln from the Six Dynasties to the Sui and Tang dynasties. Unlike the Hongzhou kiln and Xunyang city site, which were both in the Jiangxi region, the Yue kiln site is located in the Shangyu region of Zhejiang province, a straight-line distance of 600 km from the Xunyang city site alone; the distance would have been even greater if it had been over 850 km by waterway from the Yangtze River. Past research into the radiation range of Yue kiln products revealed that of the 234 burials discovered in the Jiangxi region, 11 were accompanied by celadon of Yue kiln, indicating that the Jiangxi region was within the radiation range of Yue kilns in ancient times [34]. Nevertheless, previous study [34] noted that the eleven burials in the Jiangxi region where Yue kiln celadon was discovered all belonged to the higher social classes, and all eleven of these burials belong to the Jin dynasty (266–420 A.D), which indicates that Yue kiln celadon was a symbol of the status of the higher social classes in the Jiangxi region during the Jin dynasty.

The highest social class burial discovered in the Yutu Hill burial complex at the Xunyang City site is a general's tomb, but from the Western Han period when the Yue kiln had not appeared; the burial objects in M12 and M19 indicate that their owners were of a more common status, but their burial dates are from the Sixth Dynasty to the Sui and Tang Dynasties. In conjunction with previous study, Yue kiln celadon was firstly collected by the higher social class of the Jiangxi region, by the time of the Sui and Tang dynasties, the general class of Jiangxi region had begun to collect the celadon of the Yue kiln. The production cycle of the Yue kiln began during the Eastern Han dynasty, ended during the Song dynasty, and flourished during the Tang dynasty. During the time when Yue kilns were prevalent, it is likely that all classes in the Jiangxi region collected celadon from Yue kilns. In the period following the late period of Sixth Dynasty at the Xunyang City site, Yue kiln products likely gained widespread popularity in the Jiangxi province.

Conclusion

In this study, after analysing the geochemistry of the accompanying wares from burials in the Yutu Hill burial complex at Xunyang city site, it was determined that the burial celadons of M15 and M21 tombs were of Hongzhou kiln origin, with the celadons of M15 possessing a similar REE composition to products from the Eastern Han period of the Hongzhou kiln and M21 burial celadon possessing a similar chemical composition to products from the Sui-Tang period of the Hongzhou kiln. This demonstrates that the Hongzhou kiln and Xunyang city site were connected and interacted in the Eastern Han and Sui-Tang periods.

When compared with products from other famous celadon kiln sites, the principal chemical composition of the burial celadons of M19 and M21 tombs were more similar to those of the Yue kilns, but their production techniques and external forms differed significantly from those of the Yue kilns. The geochemical analysis of the clay bodies of the accompanying celadon from burials M12 and M19 suggests that the clay originated from Zhejiang rather than Jiangxi. In addition, an internal comparison of the accompanying celadon from M12 and M19 with products from various Yue kilns reveals that their dates correspond to those of the burials. In terms of the class status of the burials M12 and M19, this paper deduces that during the late period of the Sixth Dynasty, Yue kiln products, which were only collected by the upper class in Jiangxi during the Jin dynasty, were changing to be collected by the general class.

This paper identifies a contradiction in that the accompanying celadon from M12 and M19 does not match the Yue kiln product in appearance and form, but its REE composition indicates that its clay source is nearly identical to that of the Yue kiln product. It is hoped that future research will build on this paper to investigate the specifics of the propagation of Yue kilns in Jiangxi region and whether there were undiscovered kiln sites that produced civilian wares near the Yue kiln site.