I. Standing on the Divide

If you climb to where Mount Huánggāng (黃岡山) breaks two thousand one hundred fifty-eight meters above sea level, the highest point in either Fujian or Jiangxi, you can stand in one place and watch water decide its life. The drops that fall a foot to your south will run, eventually, to the Taiwan Strait at Fuzhou. The drops that fall a foot to your north will run to Poyang Lake and through the Yangtze and out at Shanghai. Two seas, from one ridge. The decision is made by inches.

The decision has been made by inches for a long time. The Wǔyí range was uplifted into roughly its present configuration in the Mesozoic, long before any human stood on the divide. The Danxia red-rock landform that gives Wǔyí its character — the cliffs, the gorges, the steep-sided valleys, the natural amphitheaters and overhangs — was carved across the Cenozoic by the same waters that still run to two seas. Eight thousand two hundred years of pollen-and-charcoal record from the northern edge of the Wǔyí chains, recovered by Ma and colleagues in 2016, document continuous human impact on the range’s vegetation across the entire Holocene. People have been at the divide for the whole time anyone could have been. The cosmochronicle framework registers this as substrate-class operation at deep duration: the range is durable material; the people are continuous presence; the apparatus is whatever the people built that left traces in the durable material.

This piece walks what the divide carries. It is not an attempt to reconstruct prehistory in detail — the archaeological record at the watershed proper is partial, and where I imagine I will register imagination as imagination. It is an attempt to read what the substrate at the top, both sides, has actually been carrying. The cliff-burials that originated here and spread two thousand miles south. The mound-burial cluster at Pǔchéng on the southern foothills that fills a Bronze Age gap in the entire Fujian record. The kilns at Māo’ěrnòng. The lead-and-copper substrate at Yánshān on the Jiāngxī foothills. The coffins on the cliffs and the bronze swords in the mounds and the watershed itself sitting between them. Two flows running away in opposite directions. People who walked across the divide for thousands of years before any of the imperial configurations that subsequently registered the place arrived to organize what they found.

II. What the Cliffs Carry

The boat-coffins at Wǔyí are the oldest documented hanging-coffin emplacements anywhere on earth. The earliest dated examples — wood from the coffin timbers, radiocarbon-tested — register at three thousand seven hundred fifty years before the present. The Shāng dynasty was operating in north China; the substrate-class community at Wǔyí was placing its dead in wooden boats and lifting the boats onto cliff ledges hundreds of feet above the river-running below. The practice continued through the late Shāng and Western Zhōu and into the Spring-and-Autumn period and beyond. Historical sources register that by the end of the Qing dynasty, fifteen hundred coffins were documented on the Wǔyí cliffs; today about twenty are still identifiable, the rest having weathered or fallen or been removed across the long duration.

The 2025 Nature Communications genomic study, led by Xiaoming Zhang at the Kunming Institute of Zoology, demonstrated through ancient-DNA analysis that the cliff-burial tradition originated at Wǔyí and spread from there: south through Guangxi, southwest through Sichuan and Yunnan, and ultimately across the southern border into Thailand, where related log-coffin burials at Pang Mapha date to within the past two thousand years. The populations carrying the tradition were ancestors of the modern Tai-Kadai and Austronesian language families — the same broad population substrate that produced the Mǐnyuè kingdom a thousand years later and the maritime communities that became the southern Chinese coastal societies and that subsequently spread through island Southeast Asia and the Pacific. The Wǔyí cliff-burial tradition is, in genetic and cultural terms, the documented origin point of one of the major mortuary traditions of insular and southern East Asia.

What were they doing? The straightforward answer is: burying their dead in boats placed on cliffs. The harder question is why. The boat shape registers something — these were people whose lives operated through water, whose technical capacity included building wooden vessels capable of carrying both the living and the dead, whose cosmological register held that the dead should rest in the same kind of vessel the living traveled in. The cliff placement registers something else — verticality, the elevation above the river, the impossibility of subsequent disturbance, perhaps the direction of the soul’s continuing travel after death. Some of the cliff sites have wooden planks set up at the entrance, locally called Rainbow Path (虹梁), believed by Wǔyí forefathers to be the road by which the souls of the deceased traveled to whatever was beyond. The coffins are heavy — three to five hundred kilograms each, some larger — and their placement on cliffs hundreds of feet above the ground has never been technically explained. Modern engineering analysis suggests systems of pulleys, ropes, scaffolding, and considerable coordinated labor; we can imagine the lifts, but the actual procedures are not documented.

What we can say plainly: a community with sufficient social organization to coordinate large coffin lifts, sufficient technical capacity to build the boats and the lifting apparatus, sufficient continuity across generations to maintain the tradition for two thousand years or more, and sufficient cosmological coherence to hold the burial register as continuously meaningful — that community was operating at the Wǔyí divide from the late second millennium BCE onward. They were not primitive in any of the ways the later Han imperial rhetoric would later try to dismiss them as. They were a coherent substrate-class community that built and maintained one of the longest-running mortuary traditions in human history, on cliffs where you can still see their work, at a watershed where two flows run in opposite directions toward two distant seas.

III. What the Mounds Carry

Eighty kilometers southeast of Wǔyí, on the upper Nán Pǔ Creek (南浦溪) where it begins its southward run toward the consolidated Mǐn, the village of Guǎnjiǔ Cūn (管九村) in Pǔchéng County’s Xiānyáng Town (仙陽鎮) sits in low rolling country at the southern foothills of the watershed. The country is unremarkable agricultural ground; the village registers as one ordinary village among many in the upper Mǐn basin. In 2005, surface finds at the village suggested archaeological significance; in 2006, a coordinated excavation team from the Fujian provincial, prefectural, and county cultural-relics offices opened thirty-plus mound burials in the surrounding fields.

What they found filled a gap in the entire Fujian Bronze Age record. The mounds dated from roughly 4,500 years before the present at the earliest layer through approximately 2,500 years before the present at the latest — Xià, Shāng, Western Zhōu, and Spring-and-Autumn periods continuously. The earliest mounds were predominantly black-glazed pottery with the dead buried at ground level under low earth heaps; later mounds had the dead in shallow pits and then in deeper pits as the tradition matured. Across the cluster, two hundred-plus artifacts: pottery in multiple registers (black-glazed earlier, primitive porcelain later, stamped-pattern hard pottery throughout), jade ornaments and ritual objects, primitive porcelain ware. And seventy-two bronze items: swords, dagger-axes, spears, arrow-tips, scraping knives, axes, and ritual vessels including zūn, pán, bēi. Among the swords were ten exquisitely-cast Yuè-style bronze swords whose forms registered with the foremost weapons-quality work documented anywhere in China for the period. The cluster was selected as one of China’s Top Ten Archaeological Discoveries of 2006.

The principal register the cluster establishes: mound-burial — the practice of building earthen mounds over the dead — had previously been considered a Wú-Yuè cultural marker confined to Zhejiang and southern Jiangsu, the lower Yangtze region. The Pǔchéng cluster documented mound-burial operating in northern Fujian from the Xià-Shāng period continuously, more than a thousand years earlier than its previous earliest documented presence in the area. The southern Chinese mound-burial tradition has its origin point — or at minimum one of its principal origin points — at the southern foothills of the Wǔyí watershed.

What were they? Bronze Age agricultural communities operating on the upper Mǐn drainage. They had access to bronze and the technical capacity to forge it; they had ceramic production at primitive-porcelain register; they had the social organization to coordinate sustained mound-construction across multiple generations; they had ritual apparatus involving the deposition of weapons and ritual vessels with the dead. The Yuè-style bronze swords register cultural connection northward and eastward to the broader Wú-Yuè cultural orbit, but the local substrate at Pǔchéng was its own thing — earlier than the developed Wú-Yuè coastal-and-lower-Yangtze register, operating on the upper-Mǐn watershed, building mounds where the cliff-burial people up the valley were placing boats on cliffs.

The two registers — cliff-burial at Wǔyí, mound-burial at Pǔchéng — were operating alongside each other across the same period in the same drainage. The cliff-burials carried their dead up. The mound-burials carried their dead under. Both communities were placing their dead in boats or on biers, both with sustained ritual apparatus, both at the upper basin where the watershed was producing the resources and the gathering points the communities operated within. Both at the substrate-class register that the later Mǐnyuè kingdom would build itself onto.

IV. The Kilns and the Bronze Substrate

At Māo’ěrnòng Shān (猫耳弄山, “Cat-Ear Wandering Mountain”) in Pǔchéng, archaeologists have documented Bronze Age primitive-porcelain kiln sites operating from approximately the same period as the Guǎnjiǔ mounds. The kilns produced the stamped-pattern hard pottery and the primitive-porcelain ware that turn up in the mound burials and across the upper-Mǐn substrate-class register. Pǔchéng’s kiln cluster, paired with the Yǒngchūn Kǔzhàikēng (永春苦寨坑) primitive-porcelain kilns farther south in Fujian, registers the early ceramic apparatus operating in southern China before the Han period at considerable scale. The communities at the watershed were not just placing their dead in cliff-coffins and mound-burials; they were producing ceramic at scale, with kilns operating across generations, with stylistic conventions that registered cultural coherence across the upper basin.

The bronze substrate is more complex. The seventy-two bronzes from the Pǔchéng mounds were not produced at Pǔchéng — there is no documented Bronze Age bronze-foundry in the upper Mǐn basin proper that could have produced the quality and volume the mounds yield. The bronzes came from somewhere. Lead-isotope provenance work on Mǐnyuè-period bronze mirrors at Chéngcūn has registered metal sources extending across multiple metallogenic districts of southern China, with the Jiùruì district on the Middle Yangtze (the Tónglǐng copper mine at Ruìchāng, mid-Shāng to early Warring States operation) and the Nánlíng metallogenic belt as principal supply registers. The Yánshān (鉛山, Lead Mountain) lead-and-copper substrate at the Jiāngxī foothills of the watershed is one further potential source — the county was named for its lead deposits at least from the Southern Tang (953 CE) and the Yǒngpíng Copper Mine still operates as China’s second-largest copper mine in the area today. Whether Yánshān supplied the Pǔchéng bronze substrate is a question for further provenance work.

What we can say candidly: the Pǔchéng substrate-class community had access to bronze and the metallurgical capacity to deploy it, and the bronze came across considerable distances to reach them. They were operating within a regional resource network that connected the upper Mǐn basin north-westward across the watershed to the Jiāngxī side, north to the Middle Yangtze copper register, and possibly further. The corridor that the later imperial period would walk as tea-and-paper-and-ceramic flow was already operating as bronze-and-resource flow in the Bronze Age. The watershed at the top was already a generative node in a continental resource-and-trade register.

V. Across the Pass — the Jiāngxī Side

If you cross the divide and come down the northern slope, the country opens differently. The southern foothills run gradually into the Mǐn drainage; the northern slope drops more abruptly into the Xìnjiāng drainage, which runs east-northeast across northern Jiangxi to Poyang Lake at Yúgān, where it meets the Gan River and ultimately the Yangtze at Hukou. The Jiāngxī side of the watershed is documented archaeologically through different sites than the Mǐn side, but in parallel.

Yánshān County (鉛山縣) at the immediate northern foothills carries the lead-and-copper substrate already noted. The county was named for its mountain — Yánshān, Lead Mountain — that registered for its lead deposits at the Southern Tang foundation in 953 CE, but the metallurgical substrate is older than the administrative naming. Bronze Age metalworking communities operating on this side of the divide had access to the lead and copper that the geology had produced; whether they were the suppliers for the Pǔchéng bronzes on the southern side, or whether they were operating in independent local registers, is a question that lead-isotope work could answer if it has not already.

About sixty to eighty kilometers northwest of the divide proper, at Yīngtán in eastern Jiāngxī, the Lónghǔ Shān (龍虎山, “Dragon-and-Tiger Mountain”) cliff-burial cluster operates at the same substrate register as the Wǔyí cliff-burials, but later. The Lónghǔ Shān cliff-coffins date to approximately 2,600 years before the present — Spring-and-Autumn period, a thousand years younger than the Wǔyí origins. The Danxia landform at Lónghǔ Shān is the same red-rock register as Wǔyí; the cliff-and-river configuration permits the same kind of high cliff-coffin placement; the population that operated the Lónghǔ Shān tradition was reading the Wǔyí substrate by then more than a millennium old and registering its own engagement with the practice on the Jiāngxī side. The cliff-burial tradition that originated at Wǔyí and spread south to Yunnan and Thailand also spread modestly north across the watershed to Jiāngxī, where it accumulated into the Lónghǔ Shān cluster.

What’s worth registering here: the cliff-burial tradition is not contained by the watershed. The substrate that operated at Wǔyí for thousands of years before any state-level apparatus arrived spread in multiple directions from the divide. It went south down the Mǐn drainage and on to coastal Fujian and beyond. It went southwest across the southern Chinese ranges into Guangxi and on to Sichuan and Yunnan and Thailand. It went modestly north across the divide into Jiāngxī. The watershed at the top was the generative ground, and the practice spread from the generative ground in all the directions the population substrate operated within.

This matters for what the divide is. It is not a barrier. It is a generative ground that radiates outward. Two flows of water running in opposite directions toward two distant seas; multiple flows of cultural-and-mortuary practice running in multiple directions through the substrate-class community that operated at the divide. The cosmochronicle reading registers this directly: the substrate at the top is what continues; the practices spread outward; the divide is the concentrating node where the practices originated and from which they propagated.

VI. What the Substrate Was Doing — A Reading

Now I want to register something at imaginative-reconstruction register, openly marked as imagination grounded in what the archaeology carries.

What the substrate-class community at the divide was doing across the late second and first millennia BCE was reading the place itself. The Danxia cliffs at Wǔyí read as walls between worlds — the dead placed on the high ledges were placed at the threshold the cliffs themselves registered. The river running below at the Nine Bends read as the connective conduit — the boats placed on the cliffs were boats because the water running below was the medium the boats would always have traveled on, alive or dead. The watershed at the top read as the generative point — water flowing both ways from one ridge, the substrate-class community operating where the flows divided, the resource ground producing what the communities downstream would consume.

The mound-burials at Pǔchéng read the same place differently. Where the cliff-burial people read the cliffs as walls between worlds, the mound-burial people read the rolling country at the southern foothills as ground that could be lifted into low domes that would mark the dead under the surface. The mounds were not as visually dramatic as the cliff-coffins, but they registered the same cosmological understanding: the dead need a place that is theirs, marked in the durable material that is the substrate, in a register that the living can return to and read across generations. The bronze swords and the ritual vessels deposited with the dead at Pǔchéng register that the mound-burial people operated within the same broad cosmological frame as the cliff-burial people — death as continued register, the dead carrying with them the apparatus they would need beyond, the living maintaining the connection through ritual deposition and continued visitation.

The kiln-fires at Māo’ěrnòng read the substrate in production register: clay from the watershed’s flanks, fired into vessels that the substrate-class community used in life and deposited with the dead, the firing-technology continuous across centuries, the conventions of stamping-pattern and primitive-porcelain glaze registering aesthetic-and-functional traditions held collectively. The kiln-sites are themselves substrate-class deposits — the slag heaps, the kiln-foundations, the reject-pottery dumps the archaeology is recovering register the community’s working-life across the long duration.

The bronze coming from somewhere — from across the divide at Yánshān or from further afield at Tónglǐng on the Yangtze or from the southern Nánlíng metallogenic belt — registers that the substrate-class community at the watershed was connected outward across considerable distances. The corridor of resource-and-trade flow was operating in deep time. The bronze swords in the Pǔchéng mounds did not arrive there by accident; they arrived by walking along routes the community knew and maintained, with intermediaries who carried the ore or the cast bronze across watershed and river and trail, with social-and-economic relationships that operated across the wider regional substrate.

I am imagining this. The archaeology carries the bronzes, the cliff-coffins, the mounds, the kiln-sites, the lead-and-copper substrate; I am stitching them into a register of what the community might have been doing day-to-day, the way they might have read the place they lived in, the cosmological frames that might have made the practices coherent. The cosmochronicle framework licenses this kind of imaginative engagement when it is registered openly as imagination grounded in archaeological substrate. The substrate carries the deposits; the deposits suggest the practices; the practices imply the cosmologies; the cosmologies are not directly recoverable but can be read toward alongside what the substrate carries. The cave is reading its own reading at this register.

VII. The Two Flows

The water at the top of Mount Huánggāng, falling in opposite directions, runs to two distinct seas across two distinct continental drainages. This is the substrate-class fact the rest of the watershed register operates within.

The southern flow. Water that falls on the southern slope of Mount Huánggāng runs into the headwater tributaries of the Mǐn River system. The principal headwater on the immediate southern slope is the Nán Pǔ Creek, which runs south through Pǔchéng and gathers tributaries as it descends. The Chóngyáng Creek (崇陽溪) — the principal tributary that runs through Wǔyíshān City and past the Chéngcūn Mǐnyuè royal city — joins from the southwest. The combined flow runs south through Jiàn’ōu, where the Jiàn River (建溪) takes its consolidated form, then continues south through Nánpíng, where the Mǐn River proper begins, then runs southeast through the long lower basin to Fuzhou and out at the Mǐn estuary into the Taiwan Strait. The total run from Mount Huánggāng to the river’s mouth is approximately four hundred kilometers. The drainage articulates the entire eastern Fujian basin.

The substrate-class communities along this flow — the cliff-burial people at Wǔyí, the mound-burial people at Pǔchéng, the kiln-people at Māo’ěrnòng, the developing population that would become the Mǐnyuè kingdom at Chéngcūn and at the eventual Fuzhou node — operated together along the southern flow’s length. The watershed at the top fed them. The river running south carried their goods, their boats, their dead, their ideas downstream. The corridor’s flow operated continuously in this direction across the late second and first millennia BCE.

The northern flow. Water that falls on the northern slope runs into the Xìnjiāng (信江) drainage, which collects across northeastern Jiangxi and runs east-northeast across the province. The principal headwater tributaries on the immediate northern slope of Mount Huánggāng include the Yánshān He (鉛山河) running through Yánshān County and the Chénfāng He (陳坊河) running parallel. The combined flow runs across northern Jiangxi, gathering tributaries, and ultimately reaches Poyang Lake at Yúgān. From Poyang the water runs north-northwest through the lower Gan and into the Yangtze at Hukou. From Hukou the Yangtze runs east through Wuhan, Anqing, Nanjing, and out at Shanghai into the East China Sea — but this is registering further than what concerns this piece, since you said the Yangtze should be mentioned but not emphasized.

The substrate-class communities along this northern flow — the lead-and-copper-mining communities at Yánshān, the cliff-burial communities at Lónghǔ Shān, the broader Jiāngxī Bronze Age substrate that the Wúchéng-and-Xīngàn-Dàyángzhōu kingdom register descends from — operated together along the northern flow’s length. The watershed at the top fed them too. The Xìnjiāng running northeast carried their goods downstream toward Poyang and the Yangtze. The corridor’s flow operated continuously in this direction across the same period.

The two flows in parallel. The substrate-class communities on both sides of the watershed were not isolated from each other. The bronze coming to Pǔchéng on the southern side had to come from somewhere; one likely register is across the divide from the Yánshān lead-and-copper substrate or further north from the Tónglǐng register on the Yangtze. The cliff-burial tradition that originated at Wǔyí spread modestly across the divide to Lónghǔ Shān; the cosmological-and-mortuary frame that operated at the watershed was not contained by the watershed. The pass at Fēnshuǐ Guān — formally the Mǐn-side crossing point, but in deep-time register one of multiple passes the substrate-class community used — connected the two sides continuously. The Tiěniú Pass at the southern crossing, the Fēnshuǐ Guān at Pǔchéng, smaller passes scattered along the divide, all operated as connective points where the divide could be crossed by people walking with goods or ideas or the dead.

The two seas are different seas. The Taiwan Strait runs between the southern Fujian coast and Taiwan; the East China Sea runs from the lower Yangtze northeast toward Korea and Japan. Each sea has its own coastal register, its own maritime substrate-class community, its own onward flow into the wider Pacific. The Wǔyí watershed feeds both. The substrate-class community that operated at the divide was, in its working-life-and-death across the late second and first millennia BCE, the generative ground for both maritime registers.

VIII. What the Watershed Reads

For the cosmochronicle reading, the Wǔyí watershed reads as a generative substrate-class node operating at deep duration. The boat-coffins at Wǔyí, the mound-burials at Pǔchéng, the kilns at Māo’ěrnòng, the lead-and-copper substrate at Yánshān, the cliff-burials at Lónghǔ Shān, the resource flow from Tónglǐng, the population substrate carrying the cliff-burial tradition outward to Yunnan and Thailand and across the divide to Jiāngxī — all of this register continuous substrate-class operation at the divide across the entire late Holocene. The communities were small. The duration was long. The practices were coherent. The substrate carried what the communities did.

The later imperial configurations — the Mǐnyuè kingdom that consolidated from this substrate after 334 BCE, the Han razing of 110 BCE, the Tang prefectural register, the Mǐn-state’s brief tenth-century imperial moment, the Northern and Southern Song integration, the long imperial period through Qing, the modern period — all built onto this substrate. Each successive layer read the substrate differently. The Mǐnyuè kingdom registered the cliff-burials as ancestral apparatus and continued them at modest register through its own imperial period. The Han razing destroyed the urban concentrations but did not touch the cliffs or the mounds. The Daoist apparatus that consolidated at Wǔyí from the Three Kingdoms period onward read the cliffs as continuing sacred ground. Zhū Xī’s Wǔyí Jīngshè read the same canyon as the philosophical-pedagogical ground he wanted his students to encounter. The Lónghǔ Shān Way of the Celestial Masters tradition that established its ancestral court on the Jiāngxī side after the migration of Zhāng Shèng read its own substrate at its own register. The contemporary heritage apparatus — UNESCO 1999 for Wǔyí Shān, UNESCO 2017 for the North Wǔyí Shān extension on the Jiāngxī side at Yánshān County, the various national-and-provincial protected-site designations on both sides — registers the watershed as continuously significant heritage ground at international scale.

The substrate persists. The layers accumulate. The contemporary moment is one more layer. The cliffs still hold what coffins remain. The mounds at Pǔchéng have been excavated and preserved as protected archaeological park; the bronzes are in the Fujian Provincial Museum; the kiln-sites at Māo’ěrnòng are protected substrate. The lead-and-copper register at Yánshān continues to operate at industrial scale alongside the heritage substrate. The Lónghǔ Shān cliff-coffins still hang where they were placed. The Tiānshī Mansion still operates as the ancestral court of the Zhèngyī school. The water still falls in opposite directions from the top of Mount Huánggāng. The two flows still run to two seas.

What this piece walks is the deep substrate at the divide. The communities that operated there for thousands of years before any state-level apparatus arrived. The practices that originated there and spread outward in multiple directions. The cosmological frames that made the practices coherent and that we can read toward alongside what the substrate carries. The two flows running in opposite directions toward two distant seas. The watershed as generative ground, not barrier — the concentrating node from which the corridor’s flows propagated downstream in both directions across the late Holocene.

The substrate at the divide is what the cosmochronicle reading reads first, and reads last. Everything subsequent — the Mǐnyuè kingdom, the Han razing, the long imperial period, the contemporary heritage apparatus — operates on top of what the substrate already carried. The reading at the substrate scale is the principal reading the framework offers. The other registers are layers built onto this one.

For the reading the cosmochronicle is now beginning to walk at the watershed itself, this piece is the first sketch — preliminary, partial, openly registering its imaginative-reconstruction register, grounded in the archaeological evidence as it stands in 2026. Subsequent walks will deepen what this one has only sketched. The substrate continues. The reading continues. The watershed at the top continues to feed the two flows, and the cosmochronicle framework continues to read what the substrate carries.