The Women at Longhushan

A foreword on what this is

This piece reads the cliff-coffin assemblage at Longhushan in northern Jiangxi. The coffins sit in caves and crevices high above the Lushui River, on the western flank of the Wuyi range, in a country of red sandstone cliffs cut by water and weather across long time. They have been there for about twenty-five hundred years. The total at the site runs to several hundred — most of them still in their cave-positions, some of them lost to weather and looting and the cliff itself giving way over the centuries. A smaller number have been excavated under controlled conditions, and what was inside them — bodies, cloth, looms, pottery, instruments, ornaments, bamboo work, wood work — has been registered in the archaeological record.

The piece walks the assemblage. The frame is the cave's standing one: a working instrument is sited in a specific place, operates across coordinate registers, and registers what it registers at the depth the record permits. Site, agents, modes, failure conditions, textual and iconographic record. The strong / partial / faint discipline-vocabulary sits in the background and surfaces where the gauges need to register plainly.

The voice the cave aims at here is the one Edward Schafer found for The Vermilion Bird. A person who has stood in the place and seen what's there, with the science folded inside the seeing rather than nailed on top. A reader who has never opened an archaeological monograph should be able to follow what the cave is doing. A reader who knows the field should find it true and complete.

A note on collaboration. The piece has been built jointly with the AI Claude. The framework, the reading, and the substrate articulation are the cave's. The prose has been built jointly. Claude drafts. The cave revises and corrects.

The mountain and the river

Longhushan rises in the southeastern part of Jiangxi Province, about twenty kilometers southwest of Yingtan City. The name means Dragon-and-Tiger Mountain. It is not a single peak but a country of cliffs, river bends, and small forested plateaus, all carved out of red sandstone by water working downward across geological time. The formation is what geologists call Danxia — the same red-sandstone landscape that produces the famous cliffs at Wuyi, sixty kilometers to the southeast, on the other side of the watershed divide. The two mountains are geological siblings. The rock is the same. The cliffs are the same. The country is the same.

The water at Longhushan goes the other way. At Wuyi the streams gather into the Nine-Bend River and run southeast through the Min basin to the coast at Fuzhou. At Longhushan the streams gather into the Xinjiang River and run northwest into Poyang Lake and eventually into the Yangtze. The two watersheds share a divide that runs along the Wuyi range itself. Stand on the right ridge and you can almost see one water going one way and the other water going the other way.

The river at Longhushan that the cliff-coffins sit above is the Lushui — sometimes called the Luxi or the Xianshui. Lushui means roughly Reed-Water River. It runs through a country of red cliffs that rise out of the water in vertical walls, sometimes two hundred feet, sometimes more, with caves and crevices and narrow ledges scattered across the cliff faces wherever the rock weathered out differently than the rock next to it. The country has been a country of Daoist working since at least the Eastern Han: Zhang Daoling, the first of the Celestial Masters, is said to have practiced alchemy here in the second century of the common era, and the apparatus of the Celestial Masters lineage has operated at the Longhushan site continuously since then. The Daoist register sits on top of an older register: the cliff-coffins are eight or nine hundred years older than the Daoist arrival.

The cliff-coffins were noticed by outsiders only in the late 1970s. The cliff faces are too smooth to be climbed and too high to be examined from below, and the caves they sit in are sometimes invisible from the river unless you know where to look. The first systematic survey was done in 1978 and 1979. The first archaeological excavation followed in 1979 and 1980. The foundational report — Cheng Y.L. and Liu S.Z., Brief Reports on the Excavation of Guixi Cliff-Tomb in Jiangxi Province, in Wenwu (Cultural Relics) 11 (1980), one to twenty-seven, in Chinese — registers what was found in that first campaign. Subsequent campaigns have extended the record.

The specific cliff cluster where most of the coffins sit is called Xianshui Yan (仙水岩), which means roughly Immortal-Water Cliff. Eleven of the fourteen scientifically excavated tombs from the most recent registered campaign came from this cliff. The others came from Xianyan and adjacent cliffs nearby. Together the campaign yielded thirty-nine coffins, sixteen relatively complete skeletons, and more than two hundred and twenty artifacts. The artifacts include pottery, primitive celadon ware, bamboo and wooden objects, textiles, and the tools that were used to make the textiles. The dating is from the Institute of Conservation Science under the National Cultural Heritage Administration: about twenty-five hundred years on average, with Coffin Thirteen specifically registered at more than twenty-six hundred years. The window is the Spring and Autumn period sliding into the Warring States — roughly the seventh century BCE through the fifth.

Who placed them there

The people who made the cliff-coffins were a population the Chinese textual tradition came to call the Bai Yue, the Hundred Yue — the inclusive Han-period name for the various populations who lived south of the Yangtze before the southern expansions of the Qin and Han empires. The Bai Yue were not a single people. The name covers a wide range of substrate populations across what is now Zhejiang, Fujian, Jiangxi, Guangdong, Guangxi, and northern Vietnam, speaking languages mostly in the Kra-Dai and Austronesian families, with significant cultural diversity across the range. The cliff-coffin tradition is one of the registers that runs across many of these populations and pulls them into a recognizable substrate.

The genetic register of who was buried at Longhushan and the cliff-coffin sites generally has been worked at depth in the past several years. A 2020 study in iScience led by researchers at the Kunming Institute of Zoology of the Chinese Academy of Sciences, with Zhang Xiaoming as first author, analyzed mitochondrial whole genomes from forty-one human remains sampled across thirteen cliff-coffin sites in southern China. The matrilineal lines registered continuity with modern populations descended from the Bai Yue substrate. A 2025 follow-up in Nature Communications, from the same Kunming institute working with Fudan University, analyzed ancient whole genomes and confirmed direct descent from the cliff-coffin populations in the modern Bo people of Yunnan, where between forty-three and seventy-nine percent of the Bo genome traces directly to the ancient hanging-coffin practitioners. The 2025 paper also confirmed the long-standing archaeological inference: the oldest known cliff-coffin sites — about thirty-six hundred years — are in the Wuyi mountains in northern Fujian, and the practice radiated outward from there. Longhushan sits at an early western position in that radiation. The coffins are about a thousand years younger than the oldest Wuyi coffins. The substrate is the same; the watershed and the time are different.

The mtDNA evidence has implications for who specifically lies in the cliff-coffins. Matrilineal continuity registers the women's reproductive lines: the genetic substrate that connects modern populations to the cliff-coffin makers runs through the mothers. The reading is not that the women were buried preferentially in the cliffs — that question runs separately. The reading is that the women of the cliff-coffin populations carried the genetic substrate forward into the present, and that the population making and using the cliffs at Longhushan was a population in which the maternal lines were intact and continuous, sufficient to register through into modern descendants three thousand years later.

The bodies in the coffins themselves register at coordinate depths. The number of skeletons recovered in the most recent registered campaign is sixteen, distributed across thirty-nine coffins — meaning many of the coffins held no recoverable skeletal material, and some of the coffins that did hold skeletal material held multiple individuals. The pattern of multiple bodies in a single cliff-coffin is well-documented across the broader cliff-coffin tradition. Wong How Man, who has surveyed cliff-coffin sites across China and into Southeast Asia for three decades, has reported on coffins containing the bones of several individuals at once; he infers a two-stage burial practice in which the body was first interred elsewhere, allowed to decompose, and the cleaned bones were then collected and placed in the cliff coffin in a later ceremony. The bioarchaeological work at the Wangyuancun cliff-tomb in Sichuan — Han period, several centuries later than Longhushan, but in the same cliff-burial tradition — registered males, females, and children inhumed together in the same tombs, with the demographic profile suggesting that several generations of a lineage or household were placed in a single tomb. The cliff is a family container.

The work between the death and the cliff — the body-preparation, the waiting, the cleaning of the bones, the wrapping, the placement of the cleaned bones in the coffin, the ceremonies that surrounded each stage — was the substrate of human work that placed each coffin in each cave. In the comparative ethnographic record across southern China and Southeast Asia, this kind of secondary-burial work has typically run through women's hands. Reads partial on Longhushan specifically. Reads strong on the broader pattern that the cleaned-bones-in-coffins register required someone to clean the bones, and on the broader ethnographic register that this work was women's work in the populations that practiced it. The cliffs at Longhushan stand on a substrate of women's bone-working that the surviving record does not directly register but that the form of the burials requires.

The cloth and the loom

The textile recoveries from Longhushan are real, and they include both finished cloth and the tools that produced it. The recent campaign report registers textile fragments and textile tools among the more than two hundred and twenty artifacts. Among the tools registered in the broader Longhushan literature is a twill loom — a specific kind of weaving apparatus that produces a particular kind of cloth.

The distinction matters. The simplest kind of woven cloth is plain weave, in which the weft thread passes over one warp thread and under the next, then over and under again, alternating across the width of the fabric and back. Plain weave can be done on a backstrap loom, which is the simplest kind of loom: the warp threads are tied at one end to a fixed point — a wall, a post, a tree — and at the other end to a strap that goes around the weaver's lower back, and the tension of the weaver's body holds the warp threads taut. A weaver with a backstrap loom can weave plain cloth of considerable fineness, and backstrap looms were the original loom across most of Eurasia. The plant-fiber textiles registered in Chinese archaeology from the seventh and sixth millennia BCE — the hemp ropes at Hemudu, the ramie fragments at Caoheshan — are plain-weave work, done on backstrap looms or apparatus very like them.

Twill is different. In twill weave the weft passes over two or three warp threads at a time, and the pattern offsets by one warp thread on each successive pass, so that the cloth shows a diagonal ridge running across its surface. Twill is stronger than plain weave at the same thread count, holds dye differently, drapes differently, and is the structural basis of fabrics that need to take wear — denim, gabardine, herringbone, the heavy weaves of working clothing across many traditions. Twill cannot be done on a simple backstrap loom. It requires a loom with at least two heddles, the bars that lift selected warp threads up out of the way so that the weft can pass under or over them in patterns more complex than alternating single threads. A twill loom is a working machine of some sophistication.

A twill loom in a Spring-and-Autumn-period cliff-coffin assemblage in northern Jiangxi registers something substantive about the women who used it. The women at Longhushan in the seventh century BCE were not weaving the simplest kind of cloth on the simplest kind of loom. They were working at a register that required the loom to be built, maintained, set up with multiple heddle bars, threaded with patterns the weaver could keep in mind across hundreds of passes, and worked across the hours and days that the finished cloth required. The cloth itself — the surviving fragments — registers what they made. Reads partial on the specific fiber identification at Longhushan: silk, ramie, hemp, or a mixture, depending on what each individual fragment is and what the Chinese-language analytical literature has registered at depth.

The broader comparative anchor is the Mashan No. 1 Chu Tomb at Jiangling in Hubei. Mashan is dated to the fourth century BCE — a couple of hundred years later than Longhushan — and sits in the Chu state to the north and west, in the Yangtze drainage. It was excavated in 1982 and is one of the richest textile recoveries in Chinese archaeology. The coffin held a coffin cover; a silk painting; bags of utensils; costumes on wooden tomb figures; and nineteen layers of clothing and quilts around the corpse itself. The textiles included plain silks, brocades, plain and patterned gauzes, and embroidery in cross-stitch and counted stitch. The famous Mashan embroidery — dragons and phoenixes and tigers in cross-stitch on red silk — is one of the foundational artifacts of pre-imperial Chinese textile history. The level of textile working at Mashan is higher than what the Longhushan recoveries permit reading at directly, but the substrate of textile-making capacity that produced Mashan was already in place at Longhushan a few generations earlier and a few hundred miles to the southeast, and the working tradition that connects the two contexts is the same southern textile substrate that becomes the great silk-and-ramie working of the Han period and beyond.

The substrate runs deeper than either site. The earliest direct biomolecular evidence for silk in China comes from soil residues at the Neolithic site of Jiahu in Henan, dated to about eighty-five hundred years ago. Spindle whorls of stone and pottery have been recovered from sites of the fifth millennium BCE across both the Yellow River and Yangtze basins. The components of what may have been a backstrap loom are registered at sites of similar age. The hemp ropes at Hemudu in Zhejiang are seven thousand years old. The ramie textile fragments at Caoheshan in Jiangsu are six thousand. The substrate of women's textile working in what is now southern China runs back into the deep Neolithic, registers across a six-thousand-year arc from Hemudu through to the Han, and the women at Longhushan in the seventh century BCE were working in the middle of that arc, at a moment when the technical register had matured to the point where twill looms operated at the household level in cliff-coffin populations.

The thread came before the cloth. Before the loom was the spindle. Before the spindle was the harvest of the plant: ramie and hemp from cultivated fields, mulberry leaves for silkworms in baskets, the seasonal work of growing and tending and gathering the fiber. The spinning itself — the twisting of short staple fibers into continuous threads strong enough to hold tension under the loom — was a continuous working that filled the hours between other working: the spindle whorl spinning in the hand while the spinner walked, sat with children, watched fires, talked. The cloth at Longhushan registers all of this working backward through it. A few square inches of surviving ramie weave registers the months of growing the plant, weeks of retting and stripping and combing the fiber, days of spinning the thread, hours of warping the loom, hours of weaving the cloth, more hours of finishing. The cloth is the working compressed and made portable.

The cloth is also one of the things the women carried with them when they crossed the threshold of the cliff. What is less directly recoverable from the Longhushan recoveries, without engaging the primary-source analytical reports at depth, is which surviving textile fragments come from wrapping the corpse, which come from folded stacks of cloth placed as grave goods to be used or signified later, which come from clothing worn at burial, and which come from other registers entirely. The categories are different and the distinction matters. At Mashan, the nineteen layers of clothing and quilts were wound around the corpse — domestic-character textiles operating as funerary wrapping technology, with the wrapping itself a substantive ritual working that the surviving layers register at primary-source depth. At Mawangdui two centuries later, the T-shaped silk funerary banner over Lady Dai's innermost coffin was a textile purpose-built for the burial — woven, painted, and made as a soul-map for the journey beyond the cliff and the cave, with cosmological imagery (the sun and the moon, the gates of the world above, the figures of the underworld below) that does not appear on domestic-use silk and that the banner exists to carry. The two are different categories of textile working. The first is daily-life cloth pressed into funerary service. The second is funerary apparatus woven from the start. The question of which category any particular Longhushan textile fragment belongs to is the kind of question the primary-source analytical literature would carry and the reading of which would adjust the larger reading of what the assemblage is. Reads partial on the Longhushan-specific allocation. Reads strong on the broader pattern that the cliff-coffin assemblages at the Bai Yue substrate horizon include both categories of textile, and that the women at Longhushan worked at both registers.

The loom in the coffin is its own kind of question. A loom in a burial is a tool of the living placed in a death-context. Across the broader Chinese funerary record, working tools accompany the dead at varying frequencies and at varying interpretive registers — sometimes as the worker's own implement carried across, sometimes as status marker, sometimes as understood to continue serving the worker past the cliff, sometimes as ritually-significant object whose specific funerary role the surviving record does not directly register. The loom at Longhushan is in the cliff. What it was doing there, beyond the bare fact of having been placed, is the kind of question the assemblage holds open.

The cloth at Longhushan is, in significant part, the cloth the women made wrapped around the women who made it.

The pots and the celadon

The pottery from Longhushan registers at coordinate depths. The category includes ordinary hand-built and wheel-built earthenware of the kinds that fill domestic assemblages at this horizon across southern China, and it also includes what the campaign reports register as primitive celadon. The distinction matters. Celadon is a high-fired stoneware with an iron-bearing glaze that fires to a green or grey-green color under reducing conditions in the kiln. True celadon is a Yue-region invention — the famous Yue kilns at Shaoxing and Yuyao in Zhejiang, on the southeast coast — and the working tradition that produces the celebrated Tang and Song Yue ware runs back through the Han, the Warring States, and the Spring and Autumn period to a substrate of proto-celadon or primitive celadon that registers across southeast China from roughly the late Shang through the Warring States. The primitive celadon at Longhushan is one node in that substrate working.

Geometric-patterned hard-fired pottery is the characteristic Bai Yue material register. Across the southeastern coast and the inland valleys, from the Neolithic into the historic period, the Yue populations made and used a recognizable pottery with stamped geometric patterns — checkers, lozenges, zigzags, parallel lines, sometimes spirals — pressed or carved into the surface of the pot before firing. The patterns are not figurative but they are not nothing: they register a substrate of ornamental working that runs continuously across the Bai Yue range, and the pots at Longhushan participate in that substrate. The hand that pressed the pattern into the wet clay before the pot went into the fire belonged to someone — pottery-making in many premodern southern Chinese contexts ran through women's hands at the household-pottery register, with the high-production specialist kilns running through different organizations of labor, but the household pots that filled the daily working were typically women's work.

Reads strong on the pottery as substrate material at Longhushan. Reads strong on the geometric-patterned Yue register operating at the broader Bai Yue level. Reads partial on the gender-allocation of pottery-making at the specific Longhushan site, which the surviving record does not directly register. Reads partial also on the funerary-versus-domestic allocation within the recovered pottery itself — some Chinese burial assemblages at this horizon include pots made specifically for funerary deposit, sometimes never fired hard enough for daily use, sometimes shaped to register ritual purpose, alongside pots brought across from daily working. Which of the Longhushan pots fell into which category is again the kind of distinction the Chinese-language analytical literature carries at depth. The cave's working position is that the pottery in the coffin assemblage was, in substantial part, the pottery the women lived with, used, made or helped to make in the household register, and that the question of how much was purpose-built for the burial sits open and the assemblage permits reading at adjustable depths.

The ornaments and the body

The cliff-coffin assemblages across the broader tradition register a consistent range of personal-ornament categories: jade rings and earrings and ear-spools, bronze bangles and rings and anklets, beads in glass and stone and shell, and small carved ornaments in bone, antler, and tooth. The Lizhou'ao tomb in Jiangxi — Spring-and-Autumn period, in the Han-Yue transition zone, in the same province as Longhushan — yielded eleven jade pieces in a single subcomponent of the tomb. The Daxi cliff-coffin sites in Sichuan, in a different cliff-coffin tradition but at coordinate horizons, registered skeletons wearing jade rings, ear-bobs, and mussel-shell rings, with stone hoes placed under the heads of some of the dead. The broader Bronze Age Southeast Asian record registers bronze bangles, anklets, and rings as common across both male and female burials, with beads from glass and stone and shell registering participation in long trade networks that reached from the southeastern coast up into the interior.

The ornaments at Longhushan register what the women wore, and what was placed with them when they crossed into the cliff. Reads partial on specific identifications at Longhushan without access to the Chinese-language analytical literature on the individual finds. Reads strong on the broader pattern that personal ornament was a register of identity and adornment and status in the cliff-coffin populations, that women wore the ornaments in life and were buried with them, and that the trade networks the beads register reached the women at Longhushan across hundreds and thousands of miles. A glass bead in a Spring-and-Autumn-period cliff-coffin at Longhushan registers a thread of trade that may run back to the Mediterranean, where the working of glass had already been operating for a thousand years. The thread is thin but it is real.

The body itself — what the bones register about the lives the people lived — has been examined at Longhushan in the recent campaign but not, as far as Western-language scholarship has surfaced, at the depth the Wangyuancun Sichuan cliff-tomb study models for the broader cliff-coffin tradition. Stable-isotope work for diet, osteological work for occupational stress patterns, paleopathology for disease history and reproductive history, age-at-death distributions — these methodologies exist and have been applied to comparable cliff-tomb material elsewhere. Their application to the Longhushan skeletons may have been done in the Chinese-language literature and not yet surfaced in Western scholarship. The faint register on this is provisional and may strengthen.

The instrument and the music

Among the more remarkable objects in the Longhushan recoveries is a thirteen-stringed plucked musical instrument. The reports register it without elaborate description, but the instrument is consistent with what is known elsewhere about the family of zithers and zither-like instruments that were operating in southern China at the Spring-and-Autumn horizon — the se (瑟), which in its standard form had twenty-five strings, was an older and more complex instrument, and the zheng (筝), which appeared somewhat later and had fewer strings in its early forms. A thirteen-stringed instrument at Longhushan in the seventh or sixth century BCE sits in or near the early zheng register, possibly as a regional or earlier ancestor. The Mawangdui finds two centuries later included a perfectly preserved twenty-five-stringed se with silk strings intact. The instrument family was real, the working tradition was continuous, and the women buried at Longhushan lived in a culture where stringed music operated.

Who played the instrument is not directly recoverable. Across the later Chinese record, women's accomplishment in stringed-instrument playing — the qin, the zheng, the pipa — was a substantive cultural register, and the playing was often done at the domestic level rather than the court level. Whether this maps backward to Spring-and-Autumn Bai Yue culture is the kind of question that the surviving record does not directly answer. There is also the further question of whether the instrument was placed in the cliff as the personal instrument of the person whose burial it accompanied — a tool of the living crossing the threshold — or whether it was placed as part of the funerary-ceremony apparatus itself, the instrument that had been played at the death ceremony or the body-preparation working or the placement of the coffin in the cliff, and was then deposited with the burial because that was where it belonged after its working was done. Music in burial ceremony is registered across the broader Chinese funerary tradition at depth; how the Longhushan instrument participates in that register is one of the readings the assemblage holds open. The cave's reading is restrained: a stringed instrument in the assemblage registers that music was part of the cultural working at Longhushan, and that the instrument has been placed with someone whose burial the music belonged to, in one register or another, or in both. The specific player and the specific tunes are faint. The presence of music is strong.

The bamboo and the wood

The bamboo-and-wood register of the Longhushan recoveries includes the coffins themselves — single-log carvings from nanmu (phoebe) wood, sometimes shaped to suggest a boat, ranging from light coffins of a few hundred catties up to heavy coffins of more than a thousand — and a substantial inventory of bamboo and wooden objects that the campaign reports register without detailed elaboration in the Western-language summaries. Thin bamboo strips is the descriptor in the visitor-facing literature. Bamboo strips at this horizon in China include writing slips (jiǎn), which were the standard medium of textual record before paper, and they also include components of basketry, of mat-making, of cordage, of fishing apparatus, of household working broadly. The wooden objects similarly cover a substantial range: utensils, containers, small implements, the residues of household working that the wet southern climate normally destroys but that the dry caves at Longhushan have preserved.

The basketry and mat-making register is significant. Across the southern Chinese record, basket-and-mat work has typically run through women's hands. The bamboo strips at Longhushan are part of a working tradition that includes the rice winnowing baskets, the carrying baskets, the sleeping mats, the floor mats, the bamboo dippers and steamers and fish traps that constitute the substantive material substrate of household working in southern China across thousands of years. The same substrate runs through Mawangdui, where the bamboo-and-lacquer assemblage is enormous; the same substrate runs through the Hemudu Neolithic finds, where bamboo and wooden objects survive in the waterlogged conditions; the same substrate runs through to the present, where the basket-weaving working of southern Chinese rural villages is still continuous, and still in many places still women's work.

Reads strong on the bamboo-and-wood substrate at Longhushan. Reads partial on the specific identifications of individual objects, which the Chinese-language literature would carry at primary-source depth. The cave's reading registers what's there: a substantial household-working register that the women at Longhushan operated within and brought into the cliff with them.

The cliffs and the placement

How the coffins got into the cliffs is a question that has occupied the broader scholarly and popular interest for many decades. The proposed mechanisms include rope-and-pulley lowering from the top of the cliff, scaffold-and-ramp construction from the river up, and various forms of multi-stage hoisting using fixed posts driven into the cliff face. The 1989 experimental work by Lu Jingyan at Shanghai Tongji University demonstrated that lowering from above with ropes and people working in coordination was feasible at the cliff heights involved. The marks of rope wear on the rock at some cliff-coffin sites elsewhere in China support the lowering hypothesis at those sites; the marks have not been registered at Longhushan to the same degree, which has kept the specific mechanism at Longhushan a partial register.

What is strong is that the placement required organized collective working. A coffin weighing four hundred to a thousand catties, lifted or lowered into a cave fifty or eighty meters above the river, requires a working crew of at least several people coordinated in their working, with prepared ropes or scaffolding or both, and with prepared cave-positions in the cliff that may have required preliminary working to clear or shape. The working was probably done over multiple days for each coffin. The community that did this working was a community in which substantial collective effort could be organized for a funerary purpose, with the labor and the materials and the time committed to the placement of each individual coffin in its specific cave. Each coffin in its cave is the residue of a collective working that the community performed for the person inside.

Whose work the placement was, across the registers of community labor — the men climbing the cliff and lowering the ropes, the women preparing the body and the cloth and the food for the working crew, the elders directing the ceremony, the children watching from the riverbank — registers across the community register that the cliff-coffin tradition presupposes. The specific allocation of which work to whom is faint at primary-source depth and runs through inference from broader ethnographic and comparative patterns. What is strong is that the community placed the coffins together, and that women were members of that community and participated in that working.

The cliffs as a register of where the women are

Sixteen recovered skeletons across thirty-nine coffins is a small sample, and the per-coffin demographic breakdown — which coffins held one body, which held two or three, which held men, which held women, which held children, which held members of the same family — runs at the level of the Chinese-language primary-source analytical record. What can be said at the broader pattern level is this. Across cliff-coffin and cliff-tomb traditions at comparable horizons in southern China, the burials typically include women, men, and children, in mixed-sex multi-individual groupings that the demographic analyses interpret as family or lineage units. The reading of cliff-coffin populations as exclusively male, or as principally male, is not supported by the bioarchaeological evidence where the evidence has been registered.

The women at Longhushan are in the cliffs. Some of the bodies in the recovered coffins are women's bodies. The matrilineal genetic continuity registers that the population that placed the coffins included the maternal lines that descended into modern populations. The cloth in the coffins was woven by women. The pots in the coffins were used by women and in significant proportion made by women. The basketry was made by women. The clothing on the bodies was sewn by women. The body preparation between the death and the cliff was done in significant part by women. The community that did the collective working of placement included women across multiple roles, and the women's participation in that working ran through every register the assemblage permits being read.

The reading is not that the cliff-coffins are women's monuments specifically. They are family monuments — lineage containers, household containers — placed into the cliffs by the working of communities in which women and men and children participated across coordinate registers, with each person in the cliff a member of the community that placed them there. The reading is that the women's presence at Longhushan registers across every layer of the assemblage and every register of the working, and that the women's presence has been there the whole time.

What stays unreached

The Chinese-language primary-source literature on Longhushan — the Wenwu reports from 1980 forward, the subsequent Chinese-language analyses of the textile fragments, the pottery typologies, the skeletal-demographic analyses, the comparative working with the broader cliff-coffin tradition — runs at depths that this piece has not reached. The Western-language scholarly engagement with the Longhushan textile-and-women's-life material has been less developed than the engineering and genetic registers that have received more attention, and the further working at primary-source depth is among the substantial scholarly tasks that the assemblage continues to offer.

Among the questions the assemblage holds open is the funerary-versus-domestic register of the recovered objects themselves. The default reading — domestic working carried across the threshold — is probably the larger part of the truth for most cliff-coffin assemblages, and the residue of women's daily working at Longhushan is real. But the funerary-and-ritual-purpose-built category is also real, and is more developed at later sites where the evidence has been read at primary-source depth. The Mawangdui T-banner is purpose-built funerary apparatus. Lady Dai's lacquered coffins are purpose-built funerary apparatus. The bamboo-strip mortuary inventories are purpose-built. The painted wooden tomb figures are purpose-built. Some of what runs through women's hands across this tradition is body-preparation working — the cloth wrapping the body, the substances in the pots, the working that takes the body from the threshold of death to the threshold of the cliff — and that working is its own register, technical in its own ways, not reducible to the domestic-craft frame. How much of the Longhushan assemblage falls into the purpose-built funerary category, and how much of the women's working at Longhushan was body-preparation working as its own technical register rather than domestic working carried across, is a substantive question. The piece has not answered it. The assemblage has not been read at the depth the question asks for.

The specific lives of the specific women in the specific coffins are not recoverable in their particulars. Names, working biographies, family relationships, specific roles in their communities — these run at registers the surviving record does not directly carry. What the record does carry is the residue of their working: the cloth they wove, the pots they used and made, the baskets they wove, the ornaments they wore, the bodies they prepared and placed in the cliffs of their kin, their own bodies placed in cliffs by their kin in turn. The reading reaches what the record permits and stops where the working-conditions stop.

The substrate runs beneath. The matrilineal lines registered at Longhushan continue forward through the Bai Yue dispersal westward into Yunnan and southwestward into Southeast Asia, and through the modern Bo descendants, and through the various populations of southeast China and northern Vietnam whose textile-and-pottery-and-bamboo-and-household-working traditions carry the substrate that ran through the women at Longhushan twenty-five hundred years ago. The continuity is real. The working continues. The river continues to run past the cliffs. The cliffs continue to hold the coffins where the coffins have been since before the Celestial Masters arrived. The women at Longhushan are part of a working tradition that has not stopped.

References

Cheng Y.L. and Liu S.Z. 1980. "Brief Reports on the Excavation of Guixi Cliff-Tomb in Jiangxi Province." Wenwu (Cultural Relics) 11: 1–27. (In Chinese.) The foundational excavation report on Longhushan.

Gao H.W. and Wang Y.Z. 1982. "Chong'an County Wuyishan Guan Tomb Unearthed Textiles." Ethnic Studies. (In Chinese.) The comparative Wuyi textile report from the boat-coffin context across the watershed.

Hubei Provincial Museum. 1985. The No. 1 Chu Tomb of Mashan in Jiangling. Beijing: Cultural Relics Publishing House. (In Chinese.) The substantive comparative anchor for fourth-century BCE southern textile working.

Hunan Provincial Museum. 1973. The No. 1 Han Tomb of Mawangdui in Changsha. Beijing: Cultural Relics Publishing House. (In Chinese.) The Han-period extension of the southern textile substrate.

Wu Chunming. 2021. Prehistoric Maritime Frontier of Southeast China. Singapore: Springer. The substantive Western-language engagement with the Bai Yue maritime substrate.

Zhang Xiaoming, Li Chunmei, Zhou Yanan, Huang Jiahui, Yu Tengsong, Liu Xu, Shi Hong, Liu Hong, Chia Stephen, Huang Shenmin, Guo Yaozheng, Shoocongdej Rasmi, Ji Xueping, and Su Bing. 2020. "A Matrilineal Genetic Perspective of Hanging Coffin Custom in Southern China and Northern Thailand." iScience 23 (4): 101032. DOI: 10.1016/j.isci.2020.101032. (Forty-one ancient mitochondrial genomes from thirteen cliff-coffin sites; the matrilineal continuity argument.)

The 2025 follow-up paper extending the genomic working to whole-genome analysis of modern Bo descendants was published in Nature Communications on November 20, 2025, by researchers at the Kunming Institute of Zoology of the Chinese Academy of Sciences and Fudan University, confirming Wuyi as the cradle of the cliff-coffin practice and the genetic continuity of the practice into modern Bo populations.

Frontiers in Earth Science, 2023. "Bioarchaeological analysis of the human skeletal remains from cliff tomb burial of the Wangyuancun site in Leshan, Chengdu Plain, Southwest China." The comparative methodology for sex-and-age and multi-individual cliff-tomb demographics.

Sino-Platonic Papers 17 (April 1990). Heather Peters, "Tattooed Faces and Stilt Houses: Who Were the Ancient Yue?" The substantive Western-language engagement with the Bai Yue cultural register.

Schafer, Edward H. 1967. The Vermilion Bird: T'ang Images of the South. Berkeley: University of California Press. The cave's voice-register touchstone for southern-Chinese material.

Monterey, California · daveswavecave.com · May 2026