Introduction

As one of the most remarkable discoveries in Chinese archaeology in the 20th, the Sanxingdui site (4.4–2.9 ka B.P.) in the Chengdu plain (Fig. 1) of southwestern China confirms the existence of the “disappeared” ancient Shu Kingdom. The unearthed remarkable relics in the eight burial pits of the site, including exquisite large golden masks and a 2.62-m-tall bronze statue of a towering man (Fig. 1a and b), reflected it had a glorious bronze civilization, and advanced the civilization history of the Sichuan region by more than 2000 years [1, 2]. However, there are still many mysteries about this mysterious ancient Shu Kingdom, and among them is the sacrificial culture of the ancient Shu people [3]. The excavation of the pits provided a large amount of archaeological data and evidence to research the sacrificial behaviors of the ancient Shu people and explore the evolution of the ancient Shu culture (Fig. 1c), even Chinese culture [4].

Fig. 1
Fig. 1
Full size image

Photos of the exquisite cultural relics and geographical background of the Sanxingdui site. a Gold mask unearthed from K3 [14]. b The 2.62-m-tall bronze statue of a towering man unearthed from K2 [14]. c Geographical locations of the cultural regions and the sites mentioned in this paper [34]: (1) Yinxu; (2) Xiaoshuangqiao; (3) Wadian; (4) Erlitou; (5) Sanxingdui; (6) Jinsha; (7) Baodun. The yellow dotted line is the “North Route” of cultural exchange and trade between the ancient Shu Kingdom and China’s Central Plain [35]. d: Location of the Sanxingdui site

The recorded sacrificial rituals in China can be traced back to the Paleolithic Age [5]. During the Shang and Zhou dynasties, various methods such as burning, burying and sinking sacrifices were developed. In some sacrificial rituals, two or more types of methods were combined [5,6,7].

Burning sacrifice took place all over the world [8,9,10], the purpose of which was to burn sacrifices to worship ancestors and divinities by using firewood as fuel [11]. The Bible and the Epic of Gilgamesh detailed that the Israelites and Babylonians burned all or only the fat of the divided animal carcasses on the altar to worship their gods [12, 13]. The indigenous people on New Ireland Island (Latangai, Papua New Guinea) still maintain the custom of smashing and burning exquisite wooden carvings to commemorate the deceased [14]. From the Shang to Qing Dynasties, burning sacrifice was always used by Chinese emperors and nobles in their sacrificial rituals [5]. Nowadays, burning sacrifice is still used by some minorities in China [14, 15]. This kind of sacrificial rituals, burning the offerings after breaking up them, was similar to the sacrificial practices of the ancient Shu people in the Sanxingdui site with some melted and deformed fragments of artifacts buried in the pits after being crushed before burned [16]. However, a large number of the exquisite bronze and jade artifacts used by the ancient Shu people in their sacrifices indicate that the sacrificial behaviors of Sanxingdui have unique ancient Shu culture and religious characteristics, which are different from other parts of the world.

However, archaeologists speculated that the ancient Shu people held burning sacrificial rituals based on historical records and the presence of ash layers and fire-burned artifacts in the burial pits of the Sanxingdui site [17, 18], but they cannot confirm whether the burning sacrificial rituals occurred in the burial pits and whether the pits were used for sacrificial rituals or just used as trash pits to hastily bury damaged sacrificial artifacts caused by wars or other causes [8, 17,18,19,20,21,22,23,24,25]. Although ancient people's activities of using fire can be speculated through evidence obtained from historical records and the macroscopic characteristics of burnt remains [23, 26], the evidence should be verified by more scientific methods [27]. Chroma and magnetic properties provide objective indicators for exploring the ancients’ use and control of fire [28,29,30]. Because combustion can increase magnetic susceptibility (χ) and chroma redness (a*) values in sediments [31], researchers have delineated the location of the fire zone at Site No.1 of the Zhoukoudian site in China (Fig. 1c) [32] and estimated the firing temperature of artificial fired materials from Neolithic houses in Bulgaria [33]. Therefore, searching for records related to sacrificial activities in the accumulated materials of the burial pits will provide an important reference for determining the nature of the pits, interpreting such questions about the religious consciousness and behavior patterns of the ancient Shu people [20], whether there was a natural disaster, a change in the monarchy or a war [20,21,22, 24, 25].

For the above reason, in this article, we comprehensively analyze the geographical location of the pits and the accumulation characteristics of ash layers and artifacts in the pits, measured the chroma and magnetic properties of the soils near the ash layer in representative burial pits 4 and 7 (K4 and K7), and compared them with those of burnt soil debris in the ash layer in K4 and naturally deposit soil near K4. These research results can help us to clarify the purpose of the pits, understand the sacrificial activities of the ancient Shu people in the late Shang Dynasty, and explore the relationship between the Sanxingdui civilization and the Central Plains civilization (Fig. 1c).

Geographical and archaeological background of the Sanxingdui burial pits

The Sanxingdui site (104° 11′ 24″–104° 13′ 12″ E, 30° 59′ 38″–31° 06′ 33″ N) is located in the western suburbs of Guanghan in the northeastern Chengdu Plain (Fig. 1d), with a total area of 12 km2, and was first excavated in 1929 [14]. According to ancient books and archaeological evidence, it was the capital city of the ancient Shu Kingdom in the Shang Dynasty (c. 3.6–3.0 ka B.P.) [14, 36]. The site is on the southern bank of the Yazi River, which is a branch of the Jian River (a tributary of the Tuo River), and the Mamu River, another branch of the Jian River runs through the site from west to east. Floods and human activities have destroyed many of the rammed earth city walls (Fig. 2a).

Fig. 2
Fig. 2
Full size image

Location of the study area and related archaeological information [18, 44]. a Location of the study area. b Planar distribution of burial pits and soil cores. N, E, W, and S refer to the northern, eastern, western, and southern walls of K4 and K8 respectively

The eight burial pits excavated so far (K1–K8) are located in the high-platform terrain area southwest of the site (Fig. 2). The area belongs to the second terrace on the southern bank of the Mamu River [18, 37]. The upper part of the terrace sediments, with a thickness of about 3–4 m, consists mainly of brownish-yellow sandy silt, defined as Guanghan Clay deposited in the late Pleistocene [38,39,40]. The silt content of the Guanghan Clay increases gradually as the depth increases. The lower part of the terrace sediments gradually transitions to sandy gravel with increasing depth.

Except for the K5 and K6, which are products of the early Zhou Dynasty (ca. 3.0–2.8 ka) shown by unearthed cultural relics, all the pits’ depths were between 1.3 m and 2.0 m [18]. Two soil cores were taken from a location roughly 4 m west of the western wall of K4 (K4KP), and almost 5 m east of the eastern wall of K8 (K8E). These two soil cores were obtained from the same elevations as the mouths of K4 and K8, respectively (Fig. 2b). The cores indicate that the soil surrounding the pits can be roughly divided into three layers, from top to bottom (Fig. 3):

Fig. 3
Fig. 3
Full size image

Characteristics of Sanxingdui burial pits’ accumulated materials in pits 3, 4, 7 and 8 (K3, K4, K7 and K8), and of naturally deposit soil near K4(K4KP) and K8 (K8E), and sampling locations

(1) 0–20 cm: Natural sediment, brownish-yellow fine-grained sandy silt and silt, containing a few pieces of pottery shards;

(2) 20–85 cm, brown fine-grained sandy silt with high humus content. Pottery shards and burnt soil debris can be seen in the upper 20 cm thick part of this layer due to disturbance by the ancient Shu people, and the lower part is undisturbed natural sediment;

(3) 85–200 cm, naturally deposited brownish-yellow sandy silt and silt with visible iron-manganese cutans, is Guanghan Clay.

Shapes of artifacts and 14C dating results of ashes indicate that the six pits (K1–K4, K7, K8) are products of the late Shang Dynasty [41, 42]. The remnants of an artifact were scattered in the six pits, further confirming that they were produced simultaneously [43].

The morphology and accumulation characteristics of the six pits are relatively consistent, and different from K5 and K6 (both were only filled by layers of filled soil and relics from top to bottom). All six pits (K1–K4, K7, K8) are composed mainly of layers of filled soil, ashes, and relics from top to bottom, while there is an interlayer consisting of silt with a thickness of 10 to 30 cm between the layers of ash and relics in K8, which is slightly different from other five pits [18]. The filled soil layers are comprised of brown sandy silt and silt, containing pottery pieces, charcoal fragments, and burnt soil debris. The ash layers are uneven and are mainly composed of ashes and charcoal fragments from burnt gramineous plants such as bamboo, reed, rice, and millet [18]. They have no obvious boundaries with the lower relic layers (Fig. 3). Relics, including jade wares, pottery, and gold wares, dominate the bottoms of the pits [17]. The main jade wares are relatively intact, while the minor ones are fragments broken in situ. The pieces of pottery and gold wares are scattered at the bottoms of the pits, and bronze wares and fragments cover them. The large bronze wares are at the upper layers, and the small wares and fragments of bronze are at the lower layers. The bronze pieces were mainly the result of in situ crushing, and the rest formed from crushing outside the pits. Above them are the ivories scattered all over the pits. The ancient Shu people burned all the relics to varying degrees in burning sacrificial rituals [14]. The six burial pits (K1–K4, K7, K8) formed simultaneously at the Sanxingdui site all have regularly overlaying characteristics, and the relics in the pits have been burned to varying degrees, reflecting that they all likely have been used for the same purpose. All the ash layers of the pits have inclined dumping characteristics. Although ash layers are found in a single corner of K2 and K3, while an interlayer is present between the ash layer and the burnt relics layer in K8 [18]. These evidences indicate that the relics scattered all over these three pits were not burnt in the pits. However, because the ashes and charcoals scattered all over the lower parts of K1, K4 and K7, and in contact with the pits’ walls, bottoms and burnt relics layers, it is difficult to determine whether the burning sacrificial rituals occurred in the three pits only depending on the archaeological evidence. Therefore, K4 became the ideal object for our research due to its ash layer was thickest (Fig. 3) and K1 was backfilled after rescue excavation in the 1980s. In addition, we also analyzed K7 as supplements to demonstrate the reliability of the results.

Sedimentary characteristics and sampling collection of soils in and near K4 and K7

K4 is located on the north side of the excavation area of the burial pits, its mouth is approximately square (about 8.3 m2), and the elevation is 489.816 m (Figs. 2, 3). The depth of K4 is about 1.4 m (Fig. 3), and the four walls of K4 are vertical and flat. The soils near the ash layer of K4, brownish yellow sandy silt and silt, are consistent with Guanghan Clay (Fig. 3). The upper surface of K4’s ash layer inclines from the southeastern corner to the northwestern corner, the thickness of the ash layer in contact with the western wall of K4 is about 0.4 m (Fig. 3). Its components are charcoal fragments, ashes, and a small amount of irregular burnt soil debris (Fig. 3). The burning of clayey silt and silt formed the burnt soil debris. Although the surfaces of the burnt soil debris are dark red due to carbonaceous contamination during long-term buried after burnt, it has retained the morphology and grain-size characteristics of Guanghan Clay (Fig. 4). The filled soil layer in K4 covering the ash layer with an average thickness of about 85 cm (Fig. 3). The buried deep filled soil mixed pottery pieces, charcoal fragments, and burnt soil debris (Fig. 3).

Fig. 4
Fig. 4
Full size image

Characteristics of burnt soil debris, soils and Guanghan Clay. a Sampling and in-situ test locations at the burnt soil debris in K4’s ash layer. b Sampling and in-situ test locations at the western wall and bottom of K4. c Sampling and in-situ testing locations at the western wall of K7. d Sampling and in-situ testing locations at the bottom of K7. e, f Iron-manganese cutans can be seen in Guanghan Clay in core K4KP (K4KP17, K4KP20). The depth gradually increases along the arrow direction. The red cross symbols are the in-situ testing points of magnetic susceptibilities

Although compared with K4, K7 is deeper and has a thinner ash layer, the morphology and accumulation characteristics of the two pits are similar. The soils near the ash layer in K7 are comprised of brownish-yellow sandy silt and silt, and are consistent with Guanghan Clay [40].

After clearing the surface charcoal fragments and ashes, we used a portable magnetic susceptibility meter to in-situ test the magnetic susceptibilities of the burnt soil debris (K4H01-1) in K4’s ash layer and the soils near the ash layers of K4 and K7 near the ash layers (K4W-H-1, K4ST-4-1, K7NW-H-1 and K7ST-1-1). Furthermore, samples were collected from the positions of in-situ testing in K4, including burnt soil debris (K4H-01) in K4’s ash layer and the soils near the ash layer (K4W-H and K4ST-4), for laboratory magnetic property analysis. In addition, chroma samples were collected from K4 at the same positions as the in-situ magnetic susceptibilities test and from K7, including the soils near the ash layer of K7 (K7NW-H and K7ST-1). Meanwhile, samples of Guanghan Clay with depths of 105 cm and 130 cm (K4KP17 and K4KP20) were collected from K4KP for chroma and magnetic property analysis (Figs. 2, 3, 4, Table 1).

Table 1 Summary of sampling and in-situ testing point information in the Sanxingdui site
Full size table

Methods

Identifying clear evidence of humanly-controlled fire in K4 and K7 is a challenging task because the ashes and charcoals were scattered all over the lower parts of the pits and in contact with the pits’ walls, bottoms, and burnt relics layers. The characteristics of grain sizes, contents, and types of iron minerals in sediments, such as magnetite and hematite, can easily change due to the impact of human use of fire. Therefore, data of chroma and magnetic properties are used to determine whether sediments in archaeological sites were affected by the fire used by the ancients. Chroma data can be accurately measured by spectrophotometer while magnetic properties can be measured by portable magnetic susceptibility meters and laboratory magnetic property analyses. Portable magnetic susceptibility meters have the advantage of simple operation, convenient use, and high sensitivity for in situ measurements. So, we used a portable magnetic susceptibility meter to measure the burnt soil debris in ash layer of K4 and soils near the ash layer of K4 and K7. However, the accuracy of the meter’s results may be influenced by environmental magnetic fields. To demonstrate the reliability of the results, we sent the samples in K4 which tested by the meter to laboratory to measure magnetic properties. Finally, we discuss whether the ancient Shu people held burning sacrificial rituals in burial pits by integrating the data of chroma and magnetic properties.

In-situ test of magnetic susceptibilities

After the surface charcoal fragments and ashes at the test points had been cleared, a KT-10 field portable magnetic susceptibility meter (TERRAPLUS and GEORADIS s.r.o., Error is 1 × 10–6 SI, operating frequency 10 k Hz) used for in-situ measure the magnetic susceptibility (κ) of the burnt soil debris in ash layer of K4 and soils near the ash layer of K4 and K7. The value of κ reflects the total magnetic mineral concentration of sediments, dominated by ferrimagnetic minerals (e.g., magnetite, maghemite) and anti-ferromagnetic minerals (e.g., hematite, goethite).

When testing, position the instrument probe tightly against the burnt soil debris or fresh surface of the soil (avoiding cracks, pottery sherds, and burnt soil debris) after resetting the instrument, and keep the pickup coil of the instrument parallel to the fresh surface. At the same time, press the measurement button again, and the instrument display screen will show the magnetic susceptibilities of the soil. Repeat the above operation at least three times for each test point and record the results. Finally, calculate the average magnetic susceptibilities of the soil, respectively.

Laboratory measurement of chroma and magnetic properties

To reduce experimental error, we naturally dried soil samples at room temperature before sending them to the laboratory for chroma and magnetic properties tests.

Chroma measurement

Chroma measurement is an effective method for exploring the composition and content of magnetic minerals in sediments [48]. The measurements were completed at Chengdu University of Technology, using the KONICA MINOLTA CM-700d spectrophotometer (light source 10°/D65, standard deviation of chroma values ∆E*ab < 0.04). The CIELAB system, which consists of brightness (L*), redness (a*), and yellowness (b*), is used to describe the chroma of samples quantitatively [49]. The naturally exposed surfaces of the samples were trimmed to a flat surface with a moderate size before measurement, keeping the calibrated instrument measuring port close to the surface, and iron-manganese cutans of the samples should be avoided during the test. Read three sets of results for each sample and take mean values: ∆L*, ∆a*, and ∆b*.

Carbonaceous contamination will reduce the chroma values of the sample [50,51,52]. However, the contamination does not affect the ratio of the chroma values of two different wavelengths obtained from the same position of a sample surface [53,54,55]. Therefore, in this study, the ratio of ∆a*/∆b* is used to investigate the composition and relative content of magnetic minerals in the samples.

Magnetic properties measurement

Magnetic properties can effectively reflect changes in the composition, concentration and grain size of magnetic minerals in sediments caused by combustion [56, 57]. Magnetic properties were measured at South China Normal University. Fine-grained samples were packed into non-magnetic cubic boxes (V = 8 cm3) and weighed.

Values of mass susceptibilities (χ), including low- and high-frequency mass susceptibilities (χlf and χhf), were obtained as volume low- and high-frequency magnetic susceptibilities (κlf and κlf) divided with the sample density. κlf and κlf measured by Kappabridge MFK1-FA (AGICO, error <  ± 0.1%) at low (976 Hz) and high frequency (15,616 Hz) and under low fields (200 m/A). As used here, κ refers to κlf, and χ refers to χlf. The value of χ indicates the total magnetic mineral concentration of sediments. Compared to multi-domain (MD) particles, single-domain (SD) particles can obtain strong susceptibility of anhysteretic remanent magnetization (χARM) [58], so χARM is often used to measure the content of single domain particles in sediments [59, 60]. χARM was the result of the anhysteretic remanent magnetization (ARM) divided by the direct current (DC) bias (0.05 mT), the ARM was measured by a JR-6A Spinner Magnetometer (AGICO, error < 1%). The value of saturation isothermal remanent magnetization (SIRM) positively correlates with the total concentration of magnetic minerals primarily and magnetic crystal grain size secondarily, but irrelevant with superparamagnetic (SP) domains. We used an IM-10-30 Impulse Magnetizer to impart SIRM to the Z-axis for each sample at a DC field of 1T and was measured by the JR-6A Spinner Magnetometer. Both frequency-dependent magnetic susceptibility (χfd) and percentage frequency-dependent magnetic susceptibility (χfd%) values reflect the relative content of SP/SD boundary of magnetic minerals in sediments [61, 63]. The higher the values, the finer the magnetic minerals are [58,59,60]. χfd% was calculated from χfd% = [(χlf-χhf)/χlf] × 100%. Different χfd% values indicate different proportions of SP magnetic minerals in sediments: χfd% < 5%, SP magnetic minerals do not dominate the magnetic mineral assemblage; χfd% > 5%, contained a high proportion of SP magnetic minerals [61]. Isothermal remanent magnetizations (IRM-300mT) of the samples were measured under the back-field (-300 mT), and values of “hard” isothermal remanent magnetization (HIRM) were computed using the formula HIRM = (SIRM + IRM-300mT)/2 [62]. The high value of HIRM reflects the high absolute content of high coercive minerals such as hematite or goethite in the soil. In addition, S-300 was calculated using the formula S-300 = (SIRM-IRM-300mT)/(2 × SIRM) [62]. S-300 close to 1 indicates a high proportion of low coercively minerals such as magnetite and maghemite, while a low ratio (close to 0) indicates a high proportion of high coercively [59,60,61].

Results

Chroma

The chroma values and ∆a*/∆b* ratios of soils near the ash layers of K4 and K7 are similar to those of Guanghan Clay near K4 (Fig. 5).

Fig. 5
Fig. 5
Full size image

Chroma characteristics of burnt soil debris, soils and Guanghan Clay. The numbers on the horizontal axis represent the sample code: (1) K4W-H; (2) K4ST-4; (3) K7N-H; (4) K7ST-1; (5) K4KP17; (6) K4KP20; (7) K4H-01

Although the values of ∆L* and ∆b* of burnt soil debris in K4’s ash layer are lower than that of other samples due to the influence of carbonaceous contamination, the value of ∆a* of the debris is higher than other samples because of the base value of it is high (Fig. 5).

Consequently, the ∆a*/∆b* ratio of burnt soil debris is 0.74, which is significantly higher than other samples (Fig. 5).

Magnetic properties

The κ values of soil near the ash layer of K4 and the burnt soil in the ash layer, measured by KT-10, are consistent with the data obtained in the laboratory (Fig. 6), indicating that the results of KT-10 are reliable and can be compared with laboratory results. The κ values of soils near the ash layers of K4 and K7 are similar to that of Guanghan Clay near K4, significantly lower than that of the burnt soil debris (Fig. 6).

Fig. 6
Fig. 6
Full size image

Magnetic properties characteristics of burnt soil debris, soils and Guanghan Clay. The numbers on the horizontal axis represent the sample code: (1) K4W-H; (2) K4W-H-1; (3) K4ST-4; (4) K4ST-4-1; (5) K7N-H-1; (6) K7ST-1-1; (7) K4KP17; (8) K4KP20); (9) K4H-01; (10) K4H-01-1

The values of χ, χfd, χARM, SIRM, and HIRM in the soils near the ash layer of K4 are similar to those of Guanghan Clay near K4, but significantly lower than those of burnt soil debris in K4’s ash layer (Fig. 6). χfd% of burnt soil debris is above 6%, while the values of the soils near the ash layer of K4 and Guanghan Clay near K4 are below 6% (Fig. 6).

The S-300 value of burnt soil debris is approximately 0.90, close to the value of the soil from K4’s western wall, higher than the values of the soil from K4’s bottom (0.77) and Guanghan Clay near K4 (0.66) (Fig. 6).

Discussion

Indicative significance of chroma characteristics

The combustion temperature is generally closely related to the species of plants used as fuel [63, 64]. The ashes in the burial pits were mainly formed by burnt gramineous plants (such as bamboo, reed, rice and millet) growing in the Chengdu Plain in the Shang Dynasty [18, 65]. The ignition point of gramineous plants is about 300 ℃ and they are completely carbonized and gradually incinerated when the temperature reaches 400 ℃ [64, 66, 67]. In addition, serpentine jade and tremolite jade are gradually bleached after dehydration and iron oxidation reactions at 500–600 ℃ [68, 69].

The ashes in the pits are composed of carbonized and incinerated gramineous plants, and some serpentine jade and tremolite jade unearthed from the pits show partial bleaching caused by burning sacrificial rituals [70, 71], indicating the temperature of the rituals could be reached or above 500–600 ℃. Furthermore, the highest burning temperature of the open-air fire used by the ancient Shu people in the late Shang Dynasty was about 800 ℃ [72]. Consequently, we consider that the temperature of the ancient Shu people burning sacrifices could be 600–800 ℃ in the late Shang Dynasty, which is consistent with the views that the temperature of the ancients’ use of fire could be above 600 ℃ [30, 73].

Previous research suggested that lepidocrocite (γ-FeO(OH)) transforms to maghemite (γ-Fe2O3) at 300 ℃ and this may further transform to hematite at 350 ℃ [74,75,76]. In addition, goethite (α-FeO(OH)) converts to hematite (α-Fe2O3) through the dehydration process when the combustion temperature exceeds 400 ℃ [77,78,79,80]. Hematite is a typical dye mineral that can redden soil significantly [78, 81]. The value of ∆a*/∆b* can effectively reflect the ratio of hematite to goethite in soil [81,82,83,84,85].

The calculation results show that the ratios of ∆a*/∆b*of soils near the ash layers of K4 and K7 are close to that of Guanghan Clay near K4, but significantly lower than that of burnt soil debris in K4’s ash layer (Fig. 5). It indicates that the temperature of the burning sacrificial rituals can prompt a significant part of the goethite in the soil to transform into hematite, reflecting the burning temperature above 400 ℃. However, the soils near the ash layers of K4 and K7 do not appear to have been heated sufficiently to change their chroma, indicating that burning sacrificial rituals do not occur in the pits.

Indicative significance of magnetic properties

The high temperature and reductive gas environment generated in the fire process used by the ancients promoted the iron-containing silicate minerals and clay minerals in the soil to transform into fine ferromagnetic minerals, including SP and SD magnetite [46, 86]. This environment also promoted the conversion of some anti-ferromagnetic minerals in the soil into ferromagnetic minerals [46, 86]. In addition, the hematite formed by the dehydration of goethite in the soil through the fire used by the ancients is generally SP grain size [79]. The transformation of these magnetic minerals can significantly increase the values of the magnetic susceptibility of soil [31, 51, 56, 57, 87]. The κ and χ values of soils near the ash layers of K4 and K7 are similar to those of Guanghan Clay near the pits. The values are 1/5–1/11 (κ) and 1/11–1/15 (χ) of that of burnt soil debris in K4’s ash layer (Fig. 6), indicating that K4 and K7 were not affected by human use of fire.

Analysis results show that the burnt soil debris in K4’s ash layer has a high content of magnetic minerals (Fig. 6), and almost all minerals are ferromagnetic (Fig. 6). In addition, HIRM values and ∆a*/∆b* ratios show that the burnt soil debris in K4’s ash layer contains a relatively high content of imperfect anti-ferromagnetic minerals, mainly hematite (Figs. 6, 7). Ferromagnetic minerals and hematite content in soils near the ash layer of K4 are similar to those in Guanghan Clay near K4, but significantly lower than those in the burnt soil debris (Fig. 6).

Fig. 7
Fig. 7
Full size image

Biplots of magnetic properties and chroma of burnt soil debris, soils and Guanghan Clay

The results show that soils near the ash layer of K4 contain tiny SD and SP magnetic minerals, and the ratio of SP magnetic minerals to total magnetic minerals in the soil is negligible. It is similar to Guanghna Clay near K4, but contrary to burnt soil debris in the K4’s ash layer (Fig. 6). These resulted in the χ values of soils near the ash layer of K4 that were similar to Guanghan Clay but significantly lower than the value of burnt soil debris (Fig. 7).

It is considered that the magnetic mineral compositions and contents of the soils near the ash layers of K4 and K7 are similar due to the macro-sedimentary characteristics, and the values of chroma and magnetic susceptibilities, of the soils are relatively consistent with the Guanghan Clay. Therefore, the contents of fine ferromagnetic minerals and hematite in soils near the ash layers of K4 and K7 are less than that of burnt soil debris but similar to Guanghan Clay, indicating that ancient Shu people did not hold burning sacrificial rituals in the pits.

The relationship between the sacrificial cultures of the ancient Shu Kingdom and China’s Central Plain in the late Shang Dynasty

According to the chromaticity and magnetic properties of burnt soil debris in the ash layer of K4, the soils near the ash layers of K4 and K7, and the Guanghan Clay near K4, combined with the accumulation characteristics of ash and relics layers in K2, K3 and K8 [14, 18], we consider that the ashes and burnt relics were buried after the burning sacrificial rituals outside the pits formed in late Shang Dynasty. Archaeological excavations reveal that sacrificial rituals in China’s Central Plain during the Shang and Zhou Dynasties were mainly performed in the high platform terrain area [1, 5, 88]. The burial pits formed after burying sacrificial rituals (a kind of sacrificial ritual in which sacrifices are stacked in the pits in order and buried) at the Jinsha site (Fig. 1c) in the early Zhou Dynasty are also concentrated on a high platform [23, 89, 90]. The nature of the Sanxingdui burial pits is currently controversial. Some scholars believed that the pits were used to bury trash artifacts that were damaged by different reasons (such as a change in the monarchy or a war), and these pits did not have any religious purpose or sacrificial significance [15, 20]. The reason is that these burial pits were formed at the same time, but the shapes and quantities of buried artifacts in each pit vary greatly, which is different from the results of the traditional burying sacrificial rituals [43, 91]. However, these pits are concentrated in the high platform terrain area southwest of the site (Fig. 2), the shapes and overlap characteristics of the pits are regular, and they are compacted during the burial process. Moreover, the ancient Shu people buried the ivories, which were sacred sacrificial artifacts and were believed to have the function of communicating with the gods [91], on the bronze and jade artifacts in each pit during the burial process. The evidence reflects that the pits are burial places with uniform sacrificial rules, rather than the result of hasty burial, and the burial activities were devout and orderly rituals. Based on the above analysis, combined with the burial pits were formed before the ancient Shu people migrated from the Sanxingdui site due to the irresistible catastrophic events, such as natural disasters or defeat in a war, which led to the decline of the Sanxingdui civilization [92,93,94], we consider that the ancient Shu people forced to hold a series of sacrificial rituals to pray for the gods to eliminate the disasters, and the rituals went through two stages successively: (1) burnt sacrifices; (2) buried the burnt sacrifices underground. The pits were products of burying sacrificial rituals rather than buried trash artifacts [20, 22,23,24].

Many bronze and jade wares unearthed at the Sanxingdui site and the metallurgical technology of the ancient Shu Kingdom have high similarities with China’s Central Plain, such as Yinxu [14, 35] (Fig. 1c), reflecting that there was a close cultural exchange between the ancient Shu Kingdom and China’s Central Plain. Archaeologists considered that the exchange was mainly through the “North Route” [35] (Fig. 1c). In addition, the shape of some Gaobingdou (a kind of pottery food container) and other instruments unearthed at the Sanxingdui site are identical to the Erlitou site (Fig. 1c) in Yanshi, Henan Province [95]. It indicates that the cultural exchange between the two regions probably originated as early as the Xia Dynasty (4000–3600 a B.P.) or even the Longshan period (4350–3950 a B.P.) [96,97,98] (Fig. 8). Although archaeological excavation data from the Wadian site (Fig. 1c) in Yuzhou, Henan Province, reveal that rituals of burying sacrifice after burning sacrifice have already occurred in China’s Central Plain during the Longshan period (4.3–4.0 ka) [99], there are no remains of the rituals in the Baodun site (Fig. 1c) during the same period due to the preliminary and incomplete initial stage of cultural exchange [52] (Fig. 8). Subsequently, China’s Central Plain culture significantly influenced the ancient Shu culture in the Shang Dynasty (3.6–3.0 ka B.P.) with the deepening of cultural exchange [100]. At that time, the rituals of burying sacrifice after burning sacrifice occurred in China’s Central Plain [5, 15, 101]. Archaeological evidence and test data obtained by integrating sphere spectrophotometer, X-ray fluorescence (XRF), scanning electron microscope (SEM), infrared reflectance spectroscopy and Raman spectra indicated that burnt soil debris and sacrifices, which unearthed from the Xiaoshuangqaio site and Yinxu site in China’s Central Plain, experienced the rituals of burying sacrifice after burning sacrifice [102,103,104] (Fig. 1c)Thus, ancient Shu people began to hold the rituals of burying sacrifice after burning sacrifice under the influence of many factors (such as war, change of royal power, climate deterioration, etc.) in the late Shang Dynasty (~ 3.0 ka B.P.) indicates that China’s Central Plain culture probably influenced the sacrificial rituals performed by the ancient Shu people (Fig. 8). However, the sacrifices used in the sacrificial ritual, such as ivory and bronze masks, and the large scale of the burial pits, reflecting a completely different sacrificial culture compared to China’s Central Plain during the same time [5, 90, 100, 105].

Fig. 8
Fig. 8
Full size image

Compare the appearance time of the burying sacrifice after burning sacrifice between the ancient Shu Kingdom and China’s Central Plain. The arrow direction shows the process of the emergence and evolution of sacrifice. The reconstruction picture of the sacrificial scene of the ancient Shu people in the late Shang Dynasty is modified from Tang [16]. The photo of the burial pit at Sanxingdui site is cited from Ran et al. [18]

Conclusion

From the characters of the charcoal fragments, ashes and jades in the Sanxingdui buried pits, combined with the previous research results, we speculated that the temperature of burning sacrificial rituals carried by the ancient Shu people in the late Shang Dynasty could have reached 600–800 ℃. This burning temperature can lead to significant changes in the color and magnetic mineral composition of the soil. However, the Δa*/Δb* ratios and magnetic susceptibilities of the soils near the ash layers of K4 and K7, and the contents of the fine-grained (SP, SD) ferromagnetic minerals and hematite of soils in K4 and K7, are similar to those of Guanghan Clay near K4 and are significantly lower than those of burnt soil debris. Meanwhile, the ash layers in the pits present dumping characteristics, and a silty interlayer appears between the ash and relics layers in K8. These evidences indicate that the burning sacrifice did not be held in the Sanxingdui burial pits. Burial pits are distributed in the high platform terrain area in the southwest of the Sanxingdui site, and the artifacts in the pits have been stacked in a specific order. Those characteristics of burial pits indicate that the sacrificial rituals of ancient Shu people in the late Shang Dynasty included two steps: first burning in some suitable place, then burying in the pits. So, these pits were sacrificial pits and not trash pits.

During the Longshan period (4.3–4.0 ka BP) in China's Central Plain, archaeological sites have revealed the practice of burying sacrifice after burning sacrifice. However, the Sanxingdui sacrificial pits of the Shu Kingdom that emerged in the late Shang Dynasty (~ 3.0 ka BP) were unique in their large scale and the nature of their sacrifices. This suggests that despite being influenced by the Central Plain culture, the ancient Shu people of the late Shang Dynasty had their own distinct cultural and religious features.