MYS0D739RH44236·Mystery · site

Sanxingdui and its sacrificial pits

Sanxingdui lies on the Chengdu Plain northwest of Guanghan, in Sichuan. It is the largest and best-preserved Bronze Age site of the upper Yangtze, occupied from about 4400 to 3000 BP⁠[1]. In 1986 two pits packed with bronzes, gold and jade were found; between 2019 and 2022 excavators found six more. Together the eight pits have produced more than 18,000 objects, and they appear to date to the same period, about 3200 to 3000 BP, the time of the late Shang dynasty on the Central Plain⁠[1]⁠[2].

Many of the objects have no parallel elsewhere: bronze heads, masks, trees and mythical creatures, and a figure with a human head on a snake's body. Others, such as the zun and lei vessels, resemble Shang bronzes from the middle and lower Yangtze⁠[2]⁠[1]. Many were broken or burned before they were buried, and the people who made them left no writing⁠[2].

The pits are usually read as the remains of sacrifices, but that is an interpretation, and some scholars argue instead that they are the buried regalia of a destroyed polity⁠[3]. Where the bronzes were made, and whether their metal came from the same sources as Shang bronze, are being tested on the casting cores and lead isotopes⁠[1]⁠[4]. Of the 4 questions on this site, 1 are answered. The most likely answer to “Where were the Sanxingdui bronzes made, and where did their metal come from?”, not yet confirmed: heads, masks and trees were cast locally; the zun and lei vessels were made in the middle or lower Yangtze….

Key questionsset by the editors

Where were the Sanxingdui bronzes made, and where did their metal come from?Most likely: heads, masks and trees were cast locally; the zun and lei vessels were made in the middle or lower Yangtze…

Why were the Sanxingdui objects broken, burned and buried?Most likely: the objects were offerings, broken and burned in rites and buried in the sacrificial area

All 4 questions

The site

The site lies on the Chengdu Plain. The pits are in an area the excavators call the sacrificial area, within a larger Bronze Age site. No bronze-casting workshop has yet been found at Sanxingdui or anywhere else in Sichuan⁠[1]⁠[5].

The eight pits of the sacrificial areaBairuilong, 1986. CC BY-SA 4.0Artifact

Jinsha, about 30 miles (50 km) away in modern Chengdu, has similar objects and is often seen as the place the Sanxingdui people moved to after the site declined⁠[2].

Map of the areaOpen the full map

Area Region

+−

SanxingduiBronze Age walled settlement and sacrificial area

Within 1 km

Sanxingdui Museum, GuanghanMuseum

Within 100 m

Jinsha, ChengduBronze Age site

Within 1 km

MappedWithin 1 kmNatural Earth

Timeline

Timeline

Squeezed where little happened. Each mark is a published claim, by how it stands:AcceptedMost likelyStill possibleDisputedRuled out

What was found

Pits 1 and 2, excavated in 1986, held over 900 bronzes among other finds⁠[1]⁠[6]. The six new pits produced about 13,000 objects, including 1,238 bronzes, 543 gold objects and 565 jades in the counts announced in 2022, a bronze altar about 0.9 m high, a bronze box with a tortoise-shaped lid and a jade inside, apparently once wrapped in silk, and a figure with a human head and snake's body⁠[2]. The largest bronze mask is 74 cm high and 135 cm wide⁠[7]. Excavation reports for Pits 3, 7 and 8 appeared in 2024 and 2025, and the excavators published an English overview in 2025⁠[8]⁠[9].

The 1986 bronzes from Pits 1 and 2 (standing figure, heads, masks and trees)Gary Todd, 1986. CC0Artifact

Many objects in the pits were broken or burned before burial, which the excavators read as ritual treatment⁠[2].

What is known

Date. The excavators place all eight pits in one period, about 3200 to 3000 BP, on the basis of radiocarbon dating, including a published series for Pit 4⁠[1]⁠[5]⁠[10].

Where the bronzes were made. Casting cores from 39 bronzes in Pits 1 and 2 fall into groups. Cores from non-vessel objects such as heads, masks and trees contain carbonate-rich material comparable to local soils and sediments, and these objects were probably cast near Sanxingdui. Cores from the zun and lei vessels differ from local materials and match cores from casting sites in the middle and lower Yangtze⁠[1].

Metal source. Lead isotope analyses since 1995 show highly radiogenic lead in all tested Sanxingdui bronzes, vessels and non-vessels alike, the signature found in over 60% of more than 800 analysed bronzes from late Shang sites⁠[11]⁠[12]⁠[1]. Sanxingdui drew on the same metal network as the Shang even when it cast in its own style.

The explanations

Sacrifice. The usual reading is that the pits hold offerings from rites in which objects were broken, burned and buried. He Xiaoge divides the bronzes into two assemblages, one used in rites of communication with another world and one in other rites⁠[13]⁠[2].

Destruction of a polity. Some scholars read the pits as "pits of a destroyed state": the regalia and cult images of a regime, smashed and buried after conquest or overthrow⁠[3].

Ritual decommissioning. A third possibility, not argued in detail in the sources gathered here, is the disposal of a temple's furnishings when they were replaced or the centre moved; it would fit the move to Jinsha that some archaeologists infer from similar objects there⁠[2].

Natural disaster. Fan Niannian and colleagues proposed that an earthquake about 3,000 years ago dammed and diverted the river, forcing a move to Jinsha; this addresses the abandonment, not the pits⁠[14].

Published claims

Published claims and their status

Year	Who	Claimed	Status	Reason
2014	Niannian Fan and colleagues⁠[14]	An earthquake about 3,000 years ago set off landslides that dammed and diverted the river, and the people of Sanxingdui moved to Jinsha.	Still possible	Presented at a conference and reported in the press; it explains the abandonment, not the pits, and has not been tested against dated river deposits⁠[14].
2022	Sichuan Provincial Institute of Cultural Relics and Archaeology⁠[15]	Radiocarbon dating places the pits in the late Shang period, about 3,200 to 3,000 years ago.	Accepted	Consistent with the published Pit 4 series and accepted in later studies, though no model of the dates from all eight pits is in print⁠[1].
2022	Archaeologists reported by Smithsonian Magazine⁠[2]	Jinsha, about 50 km away in Chengdu, is where the Sanxingdui people moved after the site declined.	Still possible	Rests on the similarity of objects at the two sites; why and when the move happened is not established⁠[14].
2023	Ran Honglin and the Sanxingdui excavation team⁠[16]	Pits K5 and K6 were used for sacrifices, while the other pits were burial pits for ritual objects.	Still possible	The excavators' reading; whether the objects were offered, retired or destroyed in a conquest is still argued⁠[13].
2025	Huiru Lian, Xiaotian Zeng and colleagues⁠[1]	Heads, masks and trees were cast near Sanxingdui, while the zun and lei vessels were made in the middle or lower Yangtze.	Most likely	Based on 39 casting cores from Pits 1 and 2; no casting workshop has yet been found in Sichuan to confirm local production⁠[1].
2025	Huiru Lian, Xiaotian Zeng and colleagues⁠[1]	All tested Sanxingdui bronzes, vessels and non-vessels alike, contain the highly radiogenic lead found in most late Shang bronzes.	Accepted	The pattern holds in lead isotope studies since 1995, so Sanxingdui drew on the same metal network as the Shang⁠[1].
2025	He Xiaoge⁠[13]	The bronzes fall into two assemblages, one used in rites communicating with another world and one in other rites.	Still possible	An interpretive framework not yet tested against the full inventories of the new pits⁠[13].

What would settle it

Whether all eight pits were filled at once or over generations: a published Bayesian model of the radiocarbon dates from every pit, together with refits of broken fragments between pits, would show whether the burials were one event or a series. Fragments that join across pits would point to one destruction.

How the objects were destroyed: the temperatures and sequence of burning, from the bronze and soil, and whether objects were broken before or after burning, would separate a rite from a sack. A full lead isotope and alloy dataset for the new pits, released openly, would test whether different object classes drew on different metal⁠[1]⁠[4].