An Ancient Discovery in the Andes Is Leaving Archaeologists Speechless

The Ancash region of Peru's northern highlands receives few tourists. The road from Juarez descends through a narrow pass in the Cordillera Blanca, the snow-capped range that forms the spine of this part of the Andes, and then drops into the Conchucos Valley, a long, steep corridor between peaks that rarely appear on postcard maps of South America. At approximately 3,180 meters above sea level, where the Masna River meets the smaller Huecheca River, there is a flat shelf of land between two converging waterways. The rivers do not meet gently. At certain times of year, when Andean snowmelt runs fast through the high valleys, the sound of both rivers joining at this point is something close to continuous thunder. The ground itself seems to vibrate. That is where the builders chose to construct their temple, not beside the water, above it, directly over the confluence. With channels cut through the stone foundations to route the river noise into the interior of the structure itself. Chavin de Juante has been known to the outside world in some form since the 19th century, when European travelers recorded its existence and sketched its carved stone facades. The Peruvian archaeologist Julio C. Teo investigated it in the 1930s and 1940s and recognized it as one of the oldest ceremonial centers in South America. UNESCO designated it a World Heritage Site in 1985. By most conventional measures, the site had been studied. What had not been studied, what took decades of accumulated field work, acoustic physics, and laboratory chemistry to begin revealing, was what the site was actually engineered to do. The structure known as the Old Temple dates to approximately 900 BCE. That places its construction roughly contemporaneous with the early Iron Age in Europe, with the late Shang Dynasty in China, and with the period when the earliest books of the Hebrew Bible were first being composed. The people who built it left no written language. What they left is architecture. A U-shaped complex of stone platforms surrounding an open sunken plaza, with an interior that no one standing outside would suspect exists. Beneath the visible stone facades, running through the entire mass of the Old Temple, is a network of stone-walled corridors arranged in a pattern that has no architectural equivalent anywhere else in the Americas of that period. Archaeologists call them galleries. The builders left no name for them. The first thing researchers noticed, long before systematic measurement began, was that the galleries were not intuitive. They do not follow the logic of functional underground spaces, efficient pathways between points, simple grids, predictable layouts. Instead, they branch, double back, intersect at irregular angles, and create alcoves, small rectangular recesses off the main corridors, in a sequence that offers no visual orientation clues. A person walking through the galleries in darkness, which is the condition that would have prevailed during ancient use, cannot determine from the architecture alone where they are in the sequence, how deep inside the complex they have traveled, or in which direction the entrance lies. That disorientation is not an accident of construction. The evidence collected over 30 years of systematic excavation suggests it was the point. John Rick, an archaeologist and professor at Stanford University, began directing excavations at Chavin de Huanter in 1995. What Rick and his teams encountered over the following decades was a site that kept producing evidence that did not fit the initial interpretation of the galleries as simple storage or passage spaces. The stone walls, where they could be measured undisturbed, had specific proportions, narrow enough that two people could not walk side by side, low enough that taller individuals would need to adjust their posture, and lined with smaller rectangular alcoves positioned at intervals that turned out to have measurable acoustic consequences. The zigzag layout, the narrow corridors, the alcoves, each feature that initially appeared to be a construction quirk turned out to have a function that became clear only when researchers brought microphones and acoustic modeling equipment underground. In July 2001, a team excavating the Caracolas Gallery, one of the named sections within Chavin's underground network, uncovered an object that stopped work entirely. It was a conch shell, then another, then another. By the time the excavation of that particular gallery section was complete, John Rick's team had recovered 20 intact strombus galeatus shells, a marine species native to tropical Pacific waters and not found anywhere near the Peruvian highlands. Each shell had been modified in a precise and identical way. The spire filed off to create a mouthpiece. Each one was decorated with incised iconography. The complex curving visual language of Chavin art and no two were decorated identically. The wear patterns on the shells told a second story. The thick pink inner surface in several of the instruments had been worn through entirely, not from excavation damage, but from repeated use over what Rick and colleagues estimated was decades, possibly longer. The instruments had been handled so extensively that the material had been rubbed thin. Then they had been carefully wrapped in textiles, placed in the gallery, and left there. The gallery became known as the Galeria de las Caracolas, the Gallery of the Shells. The shells are now stored and studied at the Museo Nacional de Chavin, the site museum that opened in the town in 2008. They remain playable. Researchers who have handled and tested all 20 describe them as musically consistent with marine shell instruments used across the Andes from antiquity into the present, producing a deep, resonant, fundamentally low-frequency tone. What made them extraordinary in the context of Chavin was not the instruments themselves, but what their presence confirmed. Sustained, organized sonic activity inside the underground galleries over an extended period of time. The shells did not end up there by accident. They had been brought from at least 800 kilometers away, used until they showed significant wear, and then stored in a restricted underground space. The question that followed was what those instruments sounded like inside those corridors, and what effect that sound had on the people who heard it. That question took seven more years to begin answering systematically. In 2008, Rick brought a second team to Chavin. This one from Stanford's Center for Computer Research in Music and Acoustics, a unit known as CARMA. The researchers included acoustic physicists and digital signal processing specialists who had spent careers modeling the behavior of sound in architectural spaces. They brought specialized microphone arrays and measurement equipment into the galleries and began what became one of the most unusual acoustic studies ever conducted at an archaeological site. What they measured was published the same year in the Journal of the Acoustical Society of America, a peer-reviewed publication, in a paper titled simply, On the Acoustics of the Underground Galleries of Ancient Chavin de Huantar, Peru. The measurements showed several things that were individually interesting and collectively significant. First, reverberation time. The duration a sound persists in a space after its source stops. With short in the galleries, between 150 milliseconds and around one second, depending on location. In a concert hall, reverberation time might run two to three seconds. In a cathedral, longer still. Short reverberation is generally associated with spaces where sound clarity matters, where rhythmic articulation needs to be preserved. The short reverberation times in the Chavin galleries meant that percussive, rhythmic playing of the shell instruments would have been acoustically clear. Beats would not blur into each other. The second finding was more unusual. The early reflections, the first sounds bouncing off walls back to a listener's ears, were dense and energetic. Combined with a measured property called low interaural cross-correlation, meaning that what reached the left ear and what reached the right ear differed substantially, the acoustic effect produced what the researchers described as a wide sound field. A perceptual sense that sound was coming from multiple directions simultaneously, enveloping the listener rather than arriving from a locatable source. The practical consequence of low interaural cross-correlation in a narrow stone corridor is that a person standing in the gallery cannot reliably determine where a sound is coming from. Miriam Kohler, a researcher at Stanford who continued the acoustic work as the core of her doctoral dissertation and subsequent research program, tested this with living subjects. Forty-five people participated in experiments inside the galleries. Their task was to identify, while standing in various positions in the dark passageways, the direction from which a played instrument's sound originated. Results varied sharply depending on location within the gallery network. In some positions, participants accurately identified the source direction. In other positions, determined by the geometry of the corridors, the number of turns between source and listener, the presence of alcoves, their identifications were, in Kohler's own description, wildly inaccurate. The architecture, without any additional technology or manipulation, produced zones where human directional hearing became unreliable. There was a third element. The drainage canals running beneath and through the galleries, first examined in the 1970s by Peruvian archaeologist Luis Lumbereras, were not well-designed drainage channels. Lumbereras noted this in 1976 and proposed that the canals were primarily acoustic in function, that when water was introduced into the system, the routed flow through the stone passages would produce a sustained, amplified sound throughout the interior spaces. Subsequent researchers who tested this found that large volumes of water moving through the Chavin Canal system would fill the galleries with a continuous low rumble, a sound that, in scale and frequency, some researchers have compared to the vocalization of large felines, an animal with deep significance in Chavin iconography. Polished graphite mirrors, recovered in multiple excavations within the gallery system, completed the picture of what the interior environment could have been engineered to produce. Controlled shafts of reflected light directed into spaces that were otherwise in complete darkness, capable of being blocked to return those spaces to blackness. In 2017, Rixtrauel made contact with something in the sediment of a narrow underground passage that would not be fully understood for eight years. It was a tube, roughly the length and diameter of a finger, hollow, made from bone. The gallery in which he was working was a section of Chavin's underground network that had been sealed around 500 B.C., approximately at the point when the ceremonial activity at the site was declining and had not been opened again until that excavation season. The seal had held. Nine distinct stratigraphic layers were identified inside the gallery when researchers mapped its depositional sequence, evidence of material that had been placed there deliberately in organized deposits, over time, before the passage was closed. Within those layers, at floor level and in the overlying infill associated with the closure itself, the team recovered 23 artifacts. 22 were made from bone. One was made from mollusk shell. The shapes were consistent across the assemblage, tubes and spoons. No tools for cutting or preparing material. No ceramic vessels of the kind found in larger offering galleries. Only objects shaped and sized for a single specific use, inhalation. The gallery itself was approximately 7 meters long, the source describing it as a 23-foot chamber, and its physical dimensions made clear that it could hold only a small number of people at once. This was not a public space. It was not accessible from the main pilgrimage routes through the complex. Its location within the gallery network placed it in a restricted zone, separated from the circular plaza and the ceremonial areas open to the broader population of visitors. Whatever happened inside, it happened among a limited group, in a space that required deliberate navigation of the labyrinthine underground passages to reach. The bone artifacts were submitted to independent chemical and microbotanical analysis at separate institutions, as part of the protocol described in the PNAS study published on May 5, 2025. The chemical analysis used gas chromatography mass spectrometry, GCMS, to identify alkaloid compounds in the organic residues preserved inside and on the surfaces of the tubes. Six of the 23 artifacts tested positive for identifiable compounds. Those compounds were N-endimethyltryptamine, its variant 5-hydroxy-N-endimethyltryptamine, also called bufetinin, and nicotine. The first two are diagnostic markers for a plant known in the Andes as Vilca, whose seeds and leaves contain high concentrations of DMT and related alkaloids. The third is the diagnostic compound for the Nicotiana genus, wild tobacco. The microbotanical analysis confirmed the chemical findings through a different method. Starch grains from Vilca seeds and Nicotiana roots were physically present inside several of the tubes, and the starch grains showed damage consistent with dry heat, indicating that the plant material had been processed, dried, roasted, and ground before being inhaled. The animal species from which the bones were made was identified in the study as likely being Falco peregrinus, the peregrine falcon. The peregrine is a bird of documented significance in Andean iconography across multiple cultures and periods. Its winged long, hollow, and dimensionally consistent with the tubes recovered at Chavin. The choice of that particular bird, if confirmed in future zoological analysis, would place the objects within a coherent symbolic register. The predator associated with speed, precision, and altitude, its bones used to hold and deliver substances that altered the internal experience of the people who used them. The implications of the gallery's restricted access, combined with the specific chemical composition of what was used there, produced what Contreras described in the paper and subsequent interviews as a mechanism for the controlled production of hierarchical authority. The argument is not speculative in the philosophical sense, it is structural. At Chavin, the evidence shows that altered states of consciousness were not distributed equally across the population. Access to the substances, to the space where they were used, and to the acoustic and visual environment surrounding that use, was managed. The circular plaza could hold large numbers of people. The gallery where the bone tubes were found could hold a handful. The two groups experienced different versions of the same ceremonial center. One version involved the roar of water through stone canals, the sound of shell trumpets arriving from unmappable directions in the dark, and brief flashes of reflected light. The other version involved all of that, plus direct contact with psychoactive substances in a sealed underground room, in proximity to the Lanzan. That proximity is not incidental. At the center of the oldest section of the gallery network, at the point where several passages intersect, stands the Lanzan, the monolith of white granite, over four and a half meters tall, wedged floor to ceiling into the cruciform chamber of the Galeria del Lanzan. The figure carved on its surface depicts a standing being with a feline head, large upward-looking eyes, and clawed hands. Its name in Spanish refers to the stone's shape, a long tapering wedge, like a lance or a large digging stick. It does not move. It was placed in its chamber when the gallery was constructed, and it remains there because its dimensions make removal impossible without dismantling the surrounding walls. The stone was built into the temple as a permanent fixture. Nearby, measured and documented by Miriam Kohler's acoustic research team, runs a tapered stone duct approximately the same length as the Lanzan itself, open at one end to the circular plaza outside. Kohler's acoustic measurements confirm that this duct functions as a frequency-selective filter. It transmits the specific sound frequencies produced by the patu, the conch shell trumpet, while attenuating higher frequencies, including those crucial to speech clarity. The practical result is that a patutu played inside the gallery, near the Lanzan, would project its tone through the duct into the outdoor plaza in amplified and isolated form, stripped of surrounding acoustic context. To a person standing in the plaza, the Lanzan's chamber would appear to produce sound. Between 500 and 300 BCE, the system at Chavin de Huantar began to come apart. The archaeological record does not show a single event that ended it. What it shows is a sequence of gradual disassembly. The circular plaza, which had been the open ceremonial space at the center of the complex, began to be occupied by a small village. Building stones were removed from the outer structures and reused in house walls. Some carved stone tenon heads, the feline human faces that had once projected from the temple's exterior, each one thought to represent a stage in shamanistic transformation, were broken off and incorporated into domestic construction. People went on living at the site continuously for another 2,000 years, well into the 20th century, but what they built there was ordinary settlement, not ceremonial architecture. The gallery network went mostly underground and mostly sealed, some passages filled with landslide slurry from the converging rivers. Others were sealed deliberately, with the care documented in the closure deposits of the gallery excavated in 2017. The site did not end. The system that had operated within it did. The reasons for that dissolution are not settled. Scholars have proposed drought, seismic activity, shifts in regional trade routes, and internal social pressure. The Chavin model of authority was built on something inherently unstable. Access to specific experiences, sonic, chemical, visual, managed by a small group of people in a space that could not be replicated or scaled. What Chavin offered could not be distributed. It required pilgrimage to a single point in the northern highlands at 3,180 meters, passage through stone corridors in the dark, contact with objects brought from tropical waters and Amazonian forests at considerable logistical cost, and mediation by a priestly class whose authority rested on maintaining the mystery of what happened inside the restricted galleries. Once the surrounding populations began developing their own regional centers of power, with their own ceremonial architectures, their own iconographic traditions, their own trade networks, the rationale for that particular pilgrimage weakened. Chavin's influence had always traveled outward through the people who came to it and returned home carrying its imagery in ceramics, textiles, and metalwork. That dispersal eventually produced alternatives. The center became less necessary. What outlasted the center was the iconography. The Staff God, the primary deity of the Chavin visual system, depicted on the Raymondi Stela as a standing figure with feline fangs, clawed feet, and elaborate serpentine headdresses stacked in the distinctive Chavin style, appears in objects found hundreds of kilometers from the highland site. A gold crown recovered from the Lambayek Valley on the northern coast, made in the Cupisnik tradition that was roughly contemporary with Chavin, shows the same deity rendered in a coastal idiom. Cotton textiles from the southern coast, produced far from the camelid wool-producing highlands, carry the same fanged figure adapted to coastal materials. The artistic technique that defined Chavin's visual language, contour rivalry, in which the same lines can be read as two different figures depending on orientation, designed to separate those who could decode it from those who could not, appears in objects from the Paracas culture in the south and persists in the imagery of the Moche, the Tewanaku, and the wary centuries after Chavin de Juante itself had ceased to function as a pilgrimage center. The Staff God appears on the Gateway of the Sun at Tewanaku in Bolivia, a structure built roughly a thousand years after Chavin's ceremonial peak. The specific religious tradition did not survive intact, but the visual grammar it developed circulated through the Andean world long after the galleries were sealed. In 2019, a small remote controlled robot equipped with a camera was lowered into a tunnel at Chavin Tu Nero for human passage, approximately 40 centimeters in diameter. The robot's footage showed a sealed room that had not been entered since antiquity. At the center of the room was a single artifact. Three years later, in May 2022, Rick crawled through that same tunnel himself and reached the room. The artifact was a ceremonial stone bowl weighing 17 kilograms, 30 centimeters in diameter, carved with the head of an Andean condor on one side and its tail on the opposite. Rick concluded, based on the context of the deposit, that the bowl had been placed there as a votive offering when the gallery was closed, the same type of deliberate closure deposit documented in the gallery that yielded the bone tubes. The gallery was the oldest known at the site. The bowl is now documented. The full depositional context of the room is still being analyzed. Active excavation and acoustic research continue under the auspices of the Programa de Investigación Arqueológica y Conservación Chavin de Juante, authorized by Peru's Ministerio de Cultura. What the accumulated evidence from 30 years of excavation, acoustic measurement, chemical analysis, and robotic and physical exploration has produced is not a complete picture. It is a partial one, with documented boundaries. The galleries not yet accessed, the deposits not yet analyzed, the strata not yet excavated represent what remains open. What is no longer open is the prior interpretation of the site as a passive ruin, a collection of stones and carvings whose function was ceremonial in some general, unspecified sense. The evidence specifies the function in material terms. Acoustic properties measured with calibrated equipment, chemical compounds identified by mass spectrometry, stratigraphic sequences mapped by professional excavation, and living subjects tested for auditory perception inside the actual architecture. The building was engineered to do particular things to the people inside it. It did them for roughly 700 years. The record of how it worked, and what it was used for, is being extracted from the stone, the sediment, and the sealed rooms that the builders left behind.