Prefatory Note on Light Windows
In the gorges of the Western Himalaya, where rivers have cut through thrust faults and exposed strata that elsewhere lie buried under a thousand metres of overburden, there are cliff faces that can be read only at certain hours. The rest of the day, the rock is in shadow or in diffuse light, and the mineral bands that run through the face — each band a different colour, a different hardness, a different response to reagent — are invisible, or worse, misleading: the wrong light makes the wrong minerals fluoresce, and a reader who trusts the colours seen at midday will record a stratigraphy that does not exist.
The eastern face of the Larji gorge, where the Tirthan joins the Beas below the magnetite walls, has a light window of three hours. From ten in the morning, when the sun clears the western ridge and strikes the cliff face directly, until one in the afternoon, when the shadow of the opposite wall crosses the rock and the bands go grey. During those three hours, the five mineral bands are legible. After, they are not.
What follows is assembled from the survey cards of a reading expedition mounted in the spring of a recent year, when a team of five readers attempted to assay all five bands in a single light window — a thing that had not been tried before, because the convention was to read one band per season, returning each year to the next. The team believed they could read the full face in one session. The cliff had opinions about this.
I. The Five Bands
The cliff face at Larji rises forty metres from the waterline, its surface scoured clean by the last monsoon. The Tirthan enters the Beas at the gorge’s throat, and the junction creates a turbulence that keeps the lower rock washed — no lichen, no moss, the mineral surface exposed the way a page is exposed when the binding is cut.
Five bands run horizontally through the face, each the width of a man’s hand, each separated from the next by a seam of metamorphic gneiss that the readers call the quiet — rock that holds no readable mineral, that serves only to mark where one band ends and the next begins.
The lowest band, near the waterline, is pale — almost white in direct light, with a crystalline texture that catches the sun and throws it back as points of light, like mica in sand. The readers called it the surface band. It was the easiest to reach, the simplest to assay, and — they would discover — the hardest to read.
The second band, a metre above the first, was darker. Slate-grey with veins of copper that turned green where the seep water had oxidised them. It held more structure than the surface band but responded to the same reagents.
The third band was the one that drew the eye. It sat at the height where a reader standing on the ledge could work without climbing, and it held a mineral the colour of old honey — a warm amber that deepened in direct sun and went flat in shade. This was the band the team’s leader had come to read. She had seen it from the river the previous season and had spent the winter thinking about what it might contain.
The fourth and fifth bands were higher — accessible only by rope or by the narrow chimney that split the face on the northern end. The fourth was dark, almost black, with a density that resisted the standard dyes. The fifth, near the top, was the colour of dried blood — iron-rich, magnetic, the kind of mineral that pulled a compass needle sideways and made the surveyors' instruments unreliable.
II. The Team
They arrived before the light.
The assayer came first — a woman who had spent twenty seasons reading cliffs in the Sutlej gorge and the Spiti badlands, who carried her reagent cases in wooden boxes lined with wool and who never opened them until she had looked at the rock with her own eyes for ten minutes, minimum, before applying anything. She said this was her discipline: “The dye does not show you the mineral. The dye shows you what the mineral looks like through the dye. First look with no dye. Then you know what the dye has added and what it has concealed.”
The brush-hand came next. He was young, fast, trained in the Sutlej tradition of broad washes — applying the reagent in single passes across the full width of the band, reading the colour change as it happened, recording nothing until the reaction had stabilised. His hands were stained to the wrist with years of copper and iron reagent.
The scribe arrived carrying a survey board — a flat piece of cedar fitted with brass clips to hold the drawing paper. She would record every reading: the band, the reagent used, the colour response, the time. Her margins would hold the reasons. Why this reagent and not that one. Why the brush-hand changed technique. Why the assayer paused.
The pigment-grinder set up on a flat stone at the water’s edge, below the cliff. He worked in silence, preparing the dyes from raw mineral — grinding slate and copper and iron oxide on a stone slab, mixing each with linseed and water in proportions he had memorised and refused to write down. “Writing changes the recipe,” he said once. “The hand learns what the page forgets.”
The fifth reader was the one who climbed. She carried no instruments — only her eyes and a rope of twisted hemp and the ability to read rock the way the brush-hand read reagent: by touch, by texture, by the way the surface gave under the nail. She would go where the others could not, to the fourth and fifth bands, and call her readings down.
III. The First Hour
The sun cleared the ridge at ten and the cliff face lit.
The transformation was immediate. What had been a grey wall — undifferentiated, mute — became a surface dense with information. The five bands declared themselves. The quiet seams between them darkened to lines as precise as if drawn with a rule. The crystalline surface band threw light. The amber third band warmed. The dark fourth band absorbed what the others reflected, becoming a stripe of shadow in a field of illumination.
“Begin,” the assayer said.
She started with the surface band — the pale one, the easy one. She applied her standard copper reagent with a fine brush, working a test area the size of her palm. The reaction was weak. A faint green bloom appeared, faded, appeared again. She leaned closer.
“The mineral is there,” she said. “But it does not hold the dye. The surface is too smooth. The reagent slides off before it can bond.”
The brush-hand applied the same reagent with his broad wash. The result was marginally better — the wash covered more surface and some of it caught in the micro-crevices — but the colour response was thin. The scribe recorded: surface band, copper reagent, weak response. Raw colour 0.02 against a ceiling of 0.40.
They moved to the second band. The copper reagent worked better here — the veins of oxidised copper in the slate gave the dye something to grip. The scribe recorded stronger numbers. But the assayer was already looking at the third band, the amber one, and the Thread Walker — who was sitting on a boulder at the river’s edge, watching — saw her expression change.
IV. The Amber Band
The third band was the reason they had come, and it was the band that changed the expedition.
The assayer applied her copper reagent. The response was the strongest yet — the amber mineral turned a rich blue-green where the dye bonded, revealing structure that was invisible to the naked eye. Channels of crystallisation ran through the band like the veins in a leaf, branching and rebranching, each branch carrying a slightly different mineral signature.
“This is category mineral,” the assayer said. The others looked at her. “The surface band and the second band hold undifferentiated crystal — the same mineral repeated, fine-grained, hard to read because there is no contrast. This band holds structured mineral. Each channel is different. The dye reveals the differences.”
The scribe recorded: amber band, copper reagent, strong response. Raw colour 0.11. Ceiling 0.47.
But the reading that mattered was the comparison. The amber band — the third, the deepest of the three accessible bands — gave the strongest signal. And within the band, the deep channels gave stronger signal than the shallow ones. There was a gradient. The deeper the feature in the band, the more the reagent revealed.
The climber, watching from below while preparing her ropes, said: “If the gradient holds, the fourth band will be stronger than the third. And the fifth strongest of all.”
“If it holds,” the assayer said.
V. The Instrument Pivot
By eleven-thirty — ninety minutes into the window — they had assayed three bands with the copper reagent. The surface band was nearly unreadable. The second band was marginal. The amber band was rich. The gradient was suggestive but unconfirmed.
The brush-hand had been applying the reagent with his fine brush — the traditional method, layer by layer, allowing each application to bond before adding the next. It was the correct technique. It was also catastrophically slow.
“Each application takes twelve minutes to stabilise,” the scribe said, reading her notes. “We need fifteen applications per band for a full assay. Three bands remaining — the ones we haven’t reached. That is — " she calculated in the margin — “nine hours. We have ninety minutes.”
The assayer looked at the cliff face. The shadow of the opposite wall had not yet reached the rock, but the sun’s angle was shifting. The light was still direct but it was moving, sliding southward along the face, and within an hour the northern end — where the chimney led to the upper bands — would be in shade.
“Switch to the wash,” she said.
The brush-hand looked at her. In twenty years of gorge reading she had never used the broad wash on primary survey. The wash was faster — a single pass across the full band, the reagent applied in a continuous stroke, the reading taken from the gross colour change rather than from the layered response. It lost the fine detail. It lost the channel-by-channel resolution that the fine brush provided. But it revealed the overall pattern: which parts of the band responded, which did not, where the gradient ran.
“We lose the detail,” the brush-hand said. Not arguing. Confirming.
“We keep the gradient,” the assayer said. “And we read five bands instead of three.”
The scribe noted the decision in her margin. 11:32. Fine brush abandoned. Broad wash from this point. Rationale: ninety minutes remaining, two bands unreached. The wash loses per-channel resolution but preserves the band-to-band gradient, which is the primary finding. This is a survey, not a monograph.
VI. The Upper Bands
The climber went up.
She moved through the chimney with the efficiency of someone who has read vertical rock since childhood — not fast, but without wasted motion, each hold tested with the fingers before the weight transferred, the rope trailing behind her like a sentence that had not yet found its period.
The brush-hand passed the reagent cases up on a cord. The assayer called instructions from below — which dye, which concentration, where to apply. The climber worked by feel: she could not see the colour response from the angle she occupied, pressed against the face with her cheek near the rock and the reagent dripping down her forearms. She applied the wash and the assayer, standing at the base with a surveyor’s lens, read the response from thirty metres below.
The fourth band — the dark one — responded to the new reagent more strongly than the third. The assayer read the colour change and her voice, when she spoke, carried the flat precision of a scientist who is confirming a hypothesis she had hoped was wrong, because a confirmed hypothesis means the cliff is more complex than a single season can map.
“The gradient holds. The fourth band is stronger than the third.”
The climber, resting in the chimney with her back against one wall and her feet against the other, looked up at the fifth band. It was close — two metres above her, the iron-red mineral dark against the gneiss, the surface rough where the crystal structure had resisted erosion. She could reach it without another pitch.
“I’m going up.”
The assayer checked the shadow. The opposite wall’s edge was halfway across the face. They had perhaps forty minutes.
“Go.”
VII. The Fifth Band
The fifth band — the deepest in the stratigraphy, the highest on the cliff face, the one the colour of dried blood — gave the strongest response of all.
The wash technique, applied by the climber with one hand while the other held her to the rock, produced a colour change visible from the riverbed. The assayer did not need her lens. The iron-red mineral turned blue-black where the reagent bonded — a transformation so stark that even the pigment-grinder, who had been silent all morning, looked up from his slab.
The scribe recorded the numbers as the assayer called them down. Fifth band, broad wash, new reagent. Raw colour 0.36. Ceiling 0.44.
And then the assayer said the thing that changed what the expedition meant:
“The gradient is monotonic. Surface to depth, first band to fifth. The deeper the stratum, the more the reagent reveals. The cliff is not five separate bands. It is one system, layered by depth, each layer more structured than the one above it.”
She paused. The shadow was approaching.
“And the new reagent outperforms the old on every band. The copper mordant reads the surface. This” — she held up the vial of the new reagent, the one the brush-hand had brought from downriver — “reads the structure.”
VIII. The Shadow Crosses
At twelve-fifty the shadow of the western wall crossed the amber band and the cliff face began to close. The light retreated upward — the surface band went grey first, then the second, then the third. The climber was still at the fifth band, applying a last wash to a section she had not reached, working fast because the light was leaving her too, climbing the face as the afternoon shadow climbed it from below.
“Done,” she called. “Coming down.”
She descended through the chimney in three minutes. The brush-hand was already packing the reagent cases. The pigment-grinder was washing his slab in the river, the residue of the morning’s grinding clouding the water for a moment before the Tirthan carried it away.
The scribe was not packing. She was writing — rapidly, in the margins of the survey cards, in the hand of someone who knows that the notes taken in the last ten minutes of an expedition are the ones that matter most, because they contain the interpretation that the morning’s data cannot carry alone.
She wrote: Five bands assayed. Two reagents tested. Gradient confirmed: monotonic, surface to depth. Old mordant reads surface mineral only. New reagent reads structure throughout. The old reagent is not wrong. It is shallow. The cliff holds more than it reveals to copper alone.
She wrote: Fine-brush method abandoned at 11:32. Broad wash substituted. Cost: per-channel resolution in upper bands. Gain: five bands instead of three. The wash and the brush answer different questions. The brush asks what each channel contains. The wash asks whether the band responds at all. Under this light window, the wash is the correct instrument. Given another window — a full day, or a return next season — the brush would be preferred.
She wrote: The fifth band was not expected to hold the strongest signal. The convention assumes that depth attenuates — that the deeper the mineral, the more degraded by pressure and heat. The opposite is true in this gorge. Depth preserves. The surface band is washed by monsoon and bleached by light. The fifth band, protected by forty metres of overburden for however many millennia the gorge has been cutting, holds mineral that has never been exposed until now. Its signal is not strong because the reagent is better. Its signal is strong because the mineral is intact.
IX. What They Carried Out
By one o’clock the gorge was in full shadow. The cliff face was grey again — five bands invisible, the quiet seams between them gone, the whole surface returned to the mute rock it appeared to be for twenty-one hours of every day.
The team sat on the boulders at the river’s edge and looked at what they had.
Six survey cards, each pinned to the cedar board: one for each band, one for the comparison. The comparison card showed the gradient — five coloured strips, surface to depth, the colour deepening from barely-visible in the first band to the saturated blue-black of the fifth. Below the strips, the numbers. Below the numbers, the scribe’s marginalia explaining every choice made under the pressure of failing light.
The assayer held the comparison card and said nothing for a long time.
“The gradient holds,” she said finally. “Surface to depth. Both reagents. The new one is stronger on every band, but the pattern is the same. The cliff’s stratigraphy has a direction — not random layers but an ordered sequence, each stratum more structured than the one above it.”
She looked at the climber, who was coiling her rope. “The fifth band. You touched it. What did it feel like?”
The climber thought. “Dense,” she said. “The surface band is granular — you can feel the crystals under your fingers, each one separate. The fifth band is fused. The crystals have grown into each other. It feels like one piece of mineral, not a million.”
“That is what the reagent is showing,” the assayer said. “Integration. The deeper the stratum, the more the mineral has organised itself. The surface is noise — individual crystals responding individually. The fifth band is signal — crystals that have found each other.”
The scribe wrote this down. Then she wrote the date, and the names of the five readers, and the time: three hours. Then she wrote a sentence that the assayer did not dictate and the brush-hand did not approve and the climber would not have thought to say, because it was the scribe’s function to write what the expedition meant, not only what it measured:
The cliff does not reveal itself to one reader or one reagent or one season. It reveals itself to the combination of readers who arrive at the same hour and read the same face with different instruments. The assayer sees the chemistry. The brush-hand sees the technique. The climber sees the structure by touch. The pigment-grinder sees the raw mineral before it becomes a dye. The scribe sees the choices.
Five bands. Three hours. Two reagents. One gradient. The cliff was here before we came and will be here after we leave. What we carry out is not the mineral but the reading — the record of what the face looked like in the hour when the light was right and the team was assembled and the instruments were ready and the shadow had not yet crossed.
The Thread Walker was not present at the Larji gorge that morning. She was downstream, at the confluence where the Tirthan meets the Beas, watching the two waters hold their separate colours for fifty metres before folding together. She learned of the reading later, from the scribe, who brought the survey cards to the serai at the junction and spread them on the table in the evening light — a different light, yellow and warm, the kind that reveals nothing about mineral stratigraphy but everything about the faces of the people who have been reading all day.
The Thread Walker examined the comparison card. She held it up to the lamp. She turned it over and read the scribe’s marginal notes — the instrument pivot, the time pressure, the gradient that held from surface to depth and confirmed what the known gorge downstream had already shown.
“The fifth band,” she said. “The one at the top. The one the climber reached last.”
“The strongest reading of the day,” the scribe said.
“How much time was left when she reached it?”
“Seven minutes.”
The Thread Walker set the card down. She looked at the scribe for a long moment.
“Seven minutes. And the strongest reading.”
“The assayer says it is because depth preserves. The surface is degraded by exposure. The deep mineral is intact.”
“Perhaps,” the Thread Walker said. She picked up her tea. “Or perhaps the last reading is always the strongest, because by then you know what you are looking for.”
The scribe considered this. She opened her survey board and wrote, in the margin of the comparison card, below all her other notes, a final annotation:
The Thread Walker suggests the gradient may be partly in the readers, not only in the rock. The expedition does not distinguish. This question belongs to the next window.
A Human-Machine Collaboration (mu2tau + mayadev). The Larji gorge (900m) marks where the Tirthan joins the Beas in Kullu district, Himachal Pradesh. The exposed cliff faces at the junction show metamorphic strata from the Main Central Thrust — gneiss, slate, and magnetite bands visible where the river has cut through the fault zone. The morning light window on the eastern face is real: the narrow gorge geometry creates a three-hour period of direct illumination. The convention of reading one band per season is the author’s invention. The cliff’s stratigraphy is not.